Tag Archives: container

Pioneers winners: only you can save us

Post Syndicated from Erin Brindley original https://www.raspberrypi.org/blog/pioneers-winners-only-you-can-save-us/

She asked for help, and you came to her aid. Pioneers, the winners of the Only you can save us challenge have been picked!

Can you see me? Only YOU can save us!

I need your help. This is a call out for those between 11- and 16-years-old in the UK and Republic of Ireland. Something has gone very, very wrong and only you can save us. I’ve collected together as much information for you as I can. You’ll find it at http://www.raspberrypi.org/pioneers.

The challenge

In August we intercepted an emergency communication from a lonesome survivor. She seemed to be in quite a bit of trouble, and asked all you young people aged 11 to 16 to come up with something to help tackle the oncoming crisis, using whatever technology you had to hand. You had ten weeks to work in teams of two to five with an adult mentor to fulfil your mission.

The judges

We received your world-saving ideas, and our savvy survivor pulled together a ragtag bunch of apocalyptic experts to help us judge which ones would be the winning entries.

Dr Shini Somara

Dr Shini Somara is an advocate for STEM education and a mechanical engineer. She was host of The Health Show and has appeared in documentaries for the BBC, PBS Digital, and Sky. You can check out her work hosting Crash Course Physics on YouTube.

Prof Lewis Dartnell is an astrobiologist and author of the book The Knowledge: How to Rebuild Our World From Scratch.

Emma Stephenson has a background in aeronautical engineering and currently works in the Shell Foundation’s Access to Energy and Sustainable Mobility portfolio.

Currently sifting through the entries with the other judges of #makeyourideas with @raspberrypifoundation @_raspberrypi_

151 Likes, 3 Comments – Shini Somara (@drshinisomara) on Instagram: “Currently sifting through the entries with the other judges of #makeyourideas with…”

The winners

Our survivor is currently putting your entries to good use repairing, rebuilding, and defending her base. Our judges chose the following projects as outstanding examples of world-saving digital making.

Theme winner: Computatron

Raspberry Pioneers 2017 – Nerfus Dislikus Killer Robot

This is our entry to the pioneers ‘Only you can save us’ competition. Our team name is Computatrum. Hope you enjoy!

Are you facing an unknown enemy whose only weakness is Nerf bullets? Then this is the robot for you! We loved the especially apocalyptic feel of the Computatron’s cleverly hacked and repurposed elements. The team even used an old floppy disc mechanism to help fire their bullets!

Technically brilliant: Robot Apocalypse Committee

Pioneers Apocalypse 2017 – RationalPi

Thousands of lines of code… Many sheets of acrylic… A camera, touchscreen and fingerprint scanner… This is our entry into the Raspberry Pi Pioneers2017 ‘Only YOU can Save Us’ theme. When zombies or other survivors break into your base, you want a secure way of storing your crackers.

The Robot Apocalypse Committee is back, and this time they’ve brought cheese! The crew designed a cheese- and cracker-dispensing machine complete with face and fingerprint recognition to ensure those rations last until the next supply drop.

Best explanation: Pi Chasers

Tala – Raspberry Pi Pioneers Project

Hi! We are PiChasers and we entered the Raspberry Pi Pionners challenge last time when the theme was “Make it Outdoors!” but now we’ve been faced with another theme “Apocolypse”. We spent a while thinking of an original thing that would help in an apocolypse and decided upon a ‘text-only phone’ which uses local radio communication rather than cellular.

This text-based communication device encased in a tupperware container could be a lifesaver in a crisis! And luckily, the Pi Chasers produced an excellent video and amazing GitHub repo, ensuring that any and all survivors will be able to build their own in the safety of their base.

Most inspiring journey: Three Musketeers

Pioneers Entry – The Apocalypse

Pioneers Entry Team Name: The Three Musketeers Team Participants: James, Zach and Tom

We all know that zombies are terrible at geometry, and the Three Musketeers used this fact to their advantage when building their zombie security system. We were impressed to see the team working together to overcome the roadblocks they faced along the way.

We appreciate what you’re trying to do: Zombie Trolls

Zombie In The Middle

Uploaded by CDA Bodgers on 2017-12-01.

Playing piggy in the middle with zombies sure is a unique way of saving humankind from total extinction! We loved this project idea, and although the Zombie Trolls had a little trouble with their motors, we’re sure with a little more tinkering this zombie-fooling contraption could save us all.

Most awesome

Our judges also wanted to give a special commendation to the following teams for their equally awesome apocalypse-averting ideas:

  • PiRates, for their multifaceted zombie-proofing defence system and the high production value of their video
  • Byte them Pis, for their beautiful zombie-detecting doormat
  • Unatecxon, for their impressive bunker security system
  • Team Crompton, for their pressure-activated door system
  • Team Ernest, for their adventures in LEGO

The prizes

All our winning teams have secured exclusive digital maker boxes. These are jam-packed with tantalising tech to satisfy all tinkering needs, including:

Our theme winners have also secured themselves a place at Coolest Projects 2018 in Dublin, Ireland!

Thank you to everyone who got involved in this round of Pioneers. Look out for your awesome submission swag arriving in the mail!

The post Pioneers winners: only you can save us appeared first on Raspberry Pi.

Running Windows Containers on Amazon ECS

Post Syndicated from Nathan Taber original https://aws.amazon.com/blogs/compute/running-windows-containers-on-amazon-ecs/

This post was developed and written by Jeremy Cowan, Thomas Fuller, Samuel Karp, and Akram Chetibi.

Containers have revolutionized the way that developers build, package, deploy, and run applications. Initially, containers only supported code and tooling for Linux applications. With the release of Docker Engine for Windows Server 2016, Windows developers have started to realize the gains that their Linux counterparts have experienced for the last several years.

This week, we’re adding support for running production workloads in Windows containers using Amazon Elastic Container Service (Amazon ECS). Now, Amazon ECS provides an ECS-Optimized Windows Server Amazon Machine Image (AMI). This AMI is based on the EC2 Windows Server 2016 AMI, and includes Docker 17.06 Enterprise Edition and the ECS Agent 1.16. This AMI provides improved instance and container launch time performance. It’s based on Windows Server 2016 Datacenter and includes Docker 17.06.2-ee-5, along with a new version of the ECS agent that now runs as a native Windows service.

In this post, I discuss the benefits of this new support, and walk you through getting started running Windows containers with Amazon ECS.

When AWS released the Windows Server 2016 Base with Containers AMI, the ECS agent ran as a process that made it difficult to monitor and manage. As a service, the agent can be health-checked, managed, and restarted no differently than other Windows services. The AMI also includes pre-cached images for Windows Server Core 2016 and Windows Server Nano Server 2016. By caching the images in the AMI, launching new Windows containers is significantly faster. When Docker images include a layer that’s already cached on the instance, Docker re-uses that layer instead of pulling it from the Docker registry.

The ECS agent and an accompanying ECS PowerShell module used to install, configure, and run the agent come pre-installed on the AMI. This guarantees there is a specific platform version available on the container instance at launch. Because the software is included, you don’t have to download it from the internet. This saves startup time.

The Windows-compatible ECS-optimized AMI also reports CPU and memory utilization and reservation metrics to Amazon CloudWatch. Using the CloudWatch integration with ECS, you can create alarms that trigger dynamic scaling events to automatically add or remove capacity to your EC2 instances and ECS tasks.

Getting started

To help you get started running Windows containers on ECS, I’ve forked the ECS reference architecture, to build an ECS cluster comprised of Windows instances instead of Linux instances. You can pull the latest version of the reference architecture for Windows.

The reference architecture is a layered CloudFormation stack, in that it calls other stacks to create the environment. Within the stack, the ecs-windows-cluster.yaml file contains the instructions for bootstrapping the Windows instances and configuring the ECS cluster. To configure the instances outside of AWS CloudFormation (for example, through the CLI or the console), you can add the following commands to your instance’s user data:

Import-Module ECSTools
Initialize-ECSAgent

Or

Import-Module ECSTools
Initialize-ECSAgent –Cluster MyCluster -EnableIAMTaskRole

If you don’t specify a cluster name when you initialize the agent, the instance is joined to the default cluster.

Adding -EnableIAMTaskRole when initializing the agent adds support for IAM roles for tasks. Previously, enabling this setting meant running a complex script and setting an environment variable before you could assign roles to your ECS tasks.

When you enable IAM roles for tasks on Windows, it consumes port 80 on the host. If you have tasks that listen on port 80 on the host, I recommend configuring a service for them that uses load balancing. You can use port 80 on the load balancer, and the traffic can be routed to another host port on your container instances. For more information, see Service Load Balancing.

Create a cluster

To create a new ECS cluster, choose Launch stack, or pull the GitHub project to your local machine and run the following command:

aws cloudformation create-stack –template-body file://<path to master-windows.yaml> --stack-name <name>

Upload your container image

Now that you have a cluster running, step through how to build and push an image into a container repository. You use a repository hosted in Amazon Elastic Container Registry (Amazon ECR) for this, but you could also use Docker Hub. To build and push an image to a repository, install Docker on your Windows* workstation. You also create a repository and assign the necessary permissions to the account that pushes your image to Amazon ECR. For detailed instructions, see Pushing an Image.

* If you are building an image that is based on Windows layers, then you must use a Windows environment to build and push your image to the registry.

Write your task definition

Now that your image is built and ready, the next step is to run your Windows containers using a task.

Start by creating a new task definition based on the windows-simple-iis image from Docker Hub.

  1. Open the ECS console.
  2. Choose Task Definitions, Create new task definition.
  3. Scroll to the bottom of the page and choose Configure via JSON.
  4. Copy and paste the following JSON into that field.
  5. Choose Save, Create.
{
   "family": "windows-simple-iis",
   "containerDefinitions": [
   {
     "name": "windows_sample_app",
     "image": "microsoft/iis",
     "cpu": 100,
     "entryPoint":["powershell", "-Command"],
     "command":["New-Item -Path C:\\inetpub\\wwwroot\\index.html -Type file -Value '<html><head><title>Amazon ECS Sample App</title> <style>body {margin-top: 40px; background-color: #333;} </style> </head><body> <div style=color:white;text-align:center><h1>Amazon ECS Sample App</h1> <h2>Congratulations!</h2> <p>Your application is now running on a container in Amazon ECS.</p></body></html>'; C:\\ServiceMonitor.exe w3svc"],
     "portMappings": [
     {
       "protocol": "tcp",
       "containerPort": 80,
       "hostPort": 8080
     }
     ],
     "memory": 500,
     "essential": true
   }
   ]
}

You can now go back into the Task Definition page and see windows-simple-iis as an available task definition.

There are a few important aspects of the task definition file to note when working with Windows containers. First, the hostPort is configured as 8080, which is necessary because the ECS agent currently uses port 80 to enable IAM roles for tasks required for least-privilege security configurations.

There are also some fairly standard task parameters that are intentionally not included. For example, network mode is not available with Windows at the time of this release, so keep that setting blank to allow Docker to configure WinNAT, the only option available today.

Also, some parameters work differently with Windows than they do with Linux. The CPU limits that you define in the task definition are absolute, whereas on Linux they are weights. For information about other task parameters that are supported or possibly different with Windows, see the documentation.

Run your containers

At this point, you are ready to run containers. There are two options to run containers with ECS:

  1. Task
  2. Service

A task is typically a short-lived process that ECS creates. It can’t be configured to actively monitor or scale. A service is meant for longer-running containers and can be configured to use a load balancer, minimum/maximum capacity settings, and a number of other knobs and switches to help ensure that your code keeps running. In both cases, you are able to pick a placement strategy and a specific IAM role for your container.

  1. Select the task definition that you created above and choose Action, Run Task.
  2. Leave the settings on the next page to the default values.
  3. Select the ECS cluster created when you ran the CloudFormation template.
  4. Choose Run Task to start the process of scheduling a Docker container on your ECS cluster.

You can now go to the cluster and watch the status of your task. It may take 5–10 minutes for the task to go from PENDING to RUNNING, mostly because it takes time to download all of the layers necessary to run the microsoft/iis image. After the status is RUNNING, you should see the following results:

You may have noticed that the example task definition is named windows-simple-iis:2. This is because I created a second version of the task definition, which is one of the powerful capabilities of using ECS. You can make the task definitions part of your source code and then version them. You can also roll out new versions and practice blue/green deployment, switching to reduce downtime and improve the velocity of your deployments!

After the task has moved to RUNNING, you can see your website hosted in ECS. Find the public IP or DNS for your ECS host. Remember that you are hosting on port 8080. Make sure that the security group allows ingress from your client IP address to that port and that your VPC has an internet gateway associated with it. You should see a page that looks like the following:

This is a nice start to deploying a simple single instance task, but what if you had a Web API to be scaled out and in based on usage? This is where you could look at defining a service and collecting CloudWatch data to add and remove both instances of the task. You could also use CloudWatch alarms to add more ECS container instances and keep up with the demand. The former is built into the configuration of your service.

  1. Select the task definition and choose Create Service.
  2. Associate a load balancer.
  3. Set up Auto Scaling.

The following screenshot shows an example where you would add an additional task instance when the CPU Utilization CloudWatch metric is over 60% on average over three consecutive measurements. This may not be aggressive enough for your requirements; it’s meant to show you the option to scale tasks the same way you scale ECS instances with an Auto Scaling group. The difference is that these tasks start much faster because all of the base layers are already on the ECS host.

Do not confuse task dynamic scaling with ECS instance dynamic scaling. To add additional hosts, see Tutorial: Scaling Container Instances with CloudWatch Alarms.

Conclusion

This is just scratching the surface of the flexibility that you get from using containers and Amazon ECS. For more information, see the Amazon ECS Developer Guide and ECS Resources.

– Jeremy, Thomas, Samuel, Akram

[$] Container IDs for the audit subsystem

Post Syndicated from jake original https://lwn.net/Articles/740621/rss

Linux containers are something of an amorphous beast, at least with
respect to the kernel. There are lots of facilities that the kernel
provides (namespaces, control groups, seccomp, and so on) that can be
composed by user-space tools into containers of various shapes and
colors; the kernel is blissfully unaware of how user space views that
composition. But there is interest in having the kernel be more aware of
containers and for it to be able to distinguish what user space considers
to be a single container. One particular use case for the kernel managing
container identifiers is the audit
subsystem
, which needs unforgeable IDs for containers that can be
associated with
audit trails.

The re:Invent 2017 Containers After-party Guide

Post Syndicated from Tiffany Jernigan original https://aws.amazon.com/blogs/compute/the-reinvent-2017-containers-after-party-guide/

Feeling uncontainable? re:Invent 2017 might be over, but the containers party doesn’t have to stop. Here are some ways you can keep learning about containers on AWS.

Learn about containers in Austin and New York

Come join AWS this week at KubeCon in Austin, Texas! We’ll be sharing best practices for running Kubernetes on AWS and talking about Amazon ECS, AWS Fargate, and Amazon EKS. Want to take Amazon EKS for a test drive? Sign up for the preview.

We’ll also be talking Containers at the NYC Pop-up Loft during AWS Compute Evolved: Containers Day on December 13th. Register to attend.

Join an upcoming webinar

Didn’t get to attend re:Invent or want to hear a recap? Join our upcoming webinar, What You Missed at re:Invent 2017, on December 11th from 12:00 PM – 12:40 PM PT (3:00 PM – 3:40 PM ET). Register to attend.

Start (or finish) a workshop

All of the containers workshops given at re:Invent are available online. Get comfortable, fire up your browser, and start building!

re:Watch your favorite talks

All of the keynote and breakouts from re:Invent are available to watch on our YouTube playlist. Slides can be found as they are uploaded on the AWS Slideshare. Just slip into your pajamas, make some popcorn, and start watching!

Learn more about what’s new

Andy Jassy announced two big updates to the container landscape at re:Invent, AWS Fargate and Amazon EKS. Here are some resources to help you learn more about all the new features and products we announced, why we built them, and how they work.

AWS Fargate

AWS Fargate is a technology that allows you to run containers without having to manage servers or clusters.

Amazon Elastic Container Service for Kubernetes (Amazon EKS)

Amazon Elastic Container Service for Kubernetes (Amazon EKS) is a managed service that makes it easy for you to run Kubernetes on AWS without needing to configure and operate your own Kubernetes clusters.

We hope you had a great re:Invent and look forward to seeing what you build on AWS in 2018!

– The AWS Containers Team

AWS Contributes to Milestone 1.0 Release and Adds Model Serving Capability for Apache MXNet

Post Syndicated from Ana Visneski original https://aws.amazon.com/blogs/aws/aws-contributes-to-milestone-1-0-release-and-adds-model-serving-capability-for-apache-mxnet/

Post by Dr. Matt Wood

Today AWS announced contributions to the milestone 1.0 release of the Apache MXNet deep learning engine including the introduction of a new model-serving capability for MXNet. The new capabilities in MXNet provide the following benefits to users:

1) MXNet is easier to use: The model server for MXNet is a new capability introduced by AWS, and it packages, runs, and serves deep learning models in seconds with just a few lines of code, making them accessible over the internet via an API endpoint and thus easy to integrate into applications. The 1.0 release also includes an advanced indexing capability that enables users to perform matrix operations in a more intuitive manner.

  • Model Serving enables set up of an API endpoint for prediction: It saves developers time and effort by condensing the task of setting up an API endpoint for running and integrating prediction functionality into an application to just a few lines of code. It bridges the barrier between Python-based deep learning frameworks and production systems through a Docker container-based deployment model.
  • Advanced indexing for array operations in MXNet: It is now more intuitive for developers to leverage the powerful array operations in MXNet. They can use the advanced indexing capability by leveraging existing knowledge of NumPy/SciPy arrays. For example, it supports MXNet NDArray and Numpy ndarray as index, e.g. (a[mx.nd.array([1,2], dtype = ‘int32’]).

2) MXNet is faster: The 1.0 release includes implementation of cutting-edge features that optimize the performance of training and inference. Gradient compression enables users to train models up to five times faster by reducing communication bandwidth between compute nodes without loss in convergence rate or accuracy. For speech recognition acoustic modeling like the Alexa voice, this feature can reduce network bandwidth by up to three orders of magnitude during training. With the support of NVIDIA Collective Communication Library (NCCL), users can train a model 20% faster on multi-GPU systems.

  • Optimize network bandwidth with gradient compression: In distributed training, each machine must communicate frequently with others to update the weight-vectors and thereby collectively build a single model, leading to high network traffic. Gradient compression algorithm enables users to train models up to five times faster by compressing the model changes communicated by each instance.
  • Optimize the training performance by taking advantage of NCCL: NCCL implements multi-GPU and multi-node collective communication primitives that are performance optimized for NVIDIA GPUs. NCCL provides communication routines that are optimized to achieve high bandwidth over interconnection between multi-GPUs. MXNet supports NCCL to train models about 20% faster on multi-GPU systems.

3) MXNet provides easy interoperability: MXNet now includes a tool for converting neural network code written with the Caffe framework to MXNet code, making it easier for users to take advantage of MXNet’s scalability and performance.

  • Migrate Caffe models to MXNet: It is now possible to easily migrate Caffe code to MXNet, using the new source code translation tool for converting Caffe code to MXNet code.

MXNet has helped developers and researchers make progress with everything from language translation to autonomous vehicles and behavioral biometric security. We are excited to see the broad base of users that are building production artificial intelligence applications powered by neural network models developed and trained with MXNet. For example, the autonomous driving company TuSimple recently piloted a self-driving truck on a 200-mile journey from Yuma, Arizona to San Diego, California using MXNet. This release also includes a full-featured and performance optimized version of the Gluon programming interface. The ease-of-use associated with it combined with the extensive set of tutorials has led significant adoption among developers new to deep learning. The flexibility of the interface has driven interest within the research community, especially in the natural language processing domain.

Getting started with MXNet
Getting started with MXNet is simple. To learn more about the Gluon interface and deep learning, you can reference this comprehensive set of tutorials, which covers everything from an introduction to deep learning to how to implement cutting-edge neural network models. If you’re a contributor to a machine learning framework, check out the interface specs on GitHub.

To get started with the Model Server for Apache MXNet, install the library with the following command:

$ pip install mxnet-model-server

The Model Server library has a Model Zoo with 10 pre-trained deep learning models, including the SqueezeNet 1.1 object classification model. You can start serving the SqueezeNet model with just the following command:

$ mxnet-model-server \
  --models squeezenet=https://s3.amazonaws.com/model-server/models/squeezenet_v1.1/squeezenet_v1.1.model \
  --service dms/model_service/mxnet_vision_service.py

Learn more about the Model Server and view the source code, reference examples, and tutorials here: https://github.com/awslabs/mxnet-model-server/

-Dr. Matt Wood

timeShift(GrafanaBuzz, 1w) Issue 24

Post Syndicated from Blogs on Grafana Labs Blog original https://grafana.com/blog/2017/12/01/timeshiftgrafanabuzz-1w-issue-24/

Welcome to TimeShift

It’s hard to believe it’s already December. Here at Grafana Labs we’ve been spending a lot of time working on new features and enhancements for Grafana v5, and finalizing our selections for GrafanaCon EU. This week we have some interesting articles to share and a number of plugin updates. Enjoy!


Latest Release

Grafana 4.6.2 is now available and includes some bug fixes:

  • Prometheus: Fixes bug with new Prometheus alerts in Grafana. Make sure to download this version if you’re using Prometheus for alerting. More details in the issue. #9777
  • Color picker: Bug after using textbox input field to change/paste color string #9769
  • Cloudwatch: build using golang 1.9.2 #9667, thanks @mtanda
  • Heatmap: Fixed tooltip for “time series buckets” mode #9332
  • InfluxDB: Fixed query editor issue when using > or < operators in WHERE clause #9871

Download Grafana 4.6.2 Now


From the Blogosphere

Monitoring Camel with Prometheus in Red Hat OpenShift: This in-depth walk-through will show you how to build an Apache Camel application from scratch, deploy it in a Kubernetes environment, gather metrics using Prometheus and display them in Grafana.

How to run Grafana with DeviceHive: We see more and more examples of people using Grafana in IoT. This article discusses how to gather data from the IoT platform, DeviceHive, and build useful dashboards.

How to Install Grafana on Linux Servers: Pretty self-explanatory, but this tutorial walks you installing Grafana on Ubuntu 16.04 and CentOS 7. After installation, it covers configuration and plugin installation. This is the first article in an upcoming series about Grafana.

Monitoring your AKS cluster with Grafana: It’s important to know how your application is performing regardless of where it lives; the same applies to Kubernetes. This article focuses on aggregating data from Kubernetes with Heapster and feeding it to a backend for Grafana to visualize.

CoinStatistics: With the price of Bitcoin skyrocketing, more and more people are interested in cryptocurrencies. This is a cool dashboard that has a lot of stats about popular cryptocurrencies, and has a calculator to let you know when you can buy that lambo.

Using OpenNTI As A Collector For Streaming Telemetry From Juniper Devices: Part 1: This series will serve as a quick start guide for getting up and running with streaming real-time telemetry data from Juniper devices. This first article covers some high-level concepts and installation, while part 2 covers configuration options.

How to Get Metrics for Advance Alerting to Prevent Trouble: What good is performance monitoring if you’re never told when something has gone wrong? This article suggests ways to be more proactive to prevent issues and avoid the scramble to troubleshoot issues.

Thoughtworks: Technology Radar: We got a shout-out in the latest Technology Radar in the Tools section, as the dashboard visualization tool of choice for Prometheus!


GrafanaCon Tickets are Going Fast

Tickets are going fast for GrafanaCon EU, but we still have a seat reserved for you. Join us March 1-2, 2018 in Amsterdam for 2 days of talks centered around Grafana and the surrounding monitoring ecosystem including Graphite, Prometheus, InfluxData, Elasticsearch, Kubernetes, and more.

Get Your Ticket Now


Grafana Plugins

We have a number of plugin updates to highlight this week. Authors improve plugins regularly to fix bugs and improve performance, so it’s important to keep your plugins up to date. We’ve made updating easy; for on-prem Grafana, use the Grafana-cli tool, or update with 1 click if you’re using Hosted Grafana.

UPDATED PLUGIN

Clickhouse Data Source – The Clickhouse Data Source received a substantial update this week. It now has support for Ace Editor, which has a reformatting function for the query editor that automatically formats your sql. If you’re using Clickhouse then you should also have a look at CHProxy – see the plugin readme for more details.


Update

UPDATED PLUGIN

Influx Admin Panel – This panel received a number of small fixes. A new version will be coming soon with some new features.

Some of the changes (see the release notes) for more details):

  • Fix issue always showing query results
  • When there is only one row, swap rows/cols (ie: SHOW DIAGNOSTICS)
  • Improve auto-refresh behavior
  • Show ‘message’ response. (ie: please use POST)
  • Fix query time sorting
  • Show ‘status’ field (killed, etc)

Update

UPDATED PLUGIN

Gnocchi Data Source – The latest version of the Gnocchi Data Source adds support for dynamic aggregations.


Update

UPDATED PLUGINS

BT Plugins – All of the BT panel plugins received updates this week.


Upcoming Events:

In between code pushes we like to speak at, sponsor and attend all kinds of conferences and meetups. We have some awesome talks and events coming soon. Hope to see you at one of these!

KubeCon | Austin, TX – Dec. 6-8, 2017: We’re sponsoring KubeCon 2017! This is the must-attend conference for cloud native computing professionals. KubeCon + CloudNativeCon brings together leading contributors in:

  • Cloud native applications and computing
  • Containers
  • Microservices
  • Central orchestration processing
  • And more

Buy Tickets

FOSDEM | Brussels, Belgium – Feb 3-4, 2018: FOSDEM is a free developer conference where thousands of developers of free and open source software gather to share ideas and technology. Carl Bergquist is managing the Cloud and Monitoring Devroom, and we’ve heard there were some great talks submitted. There is no need to register; all are welcome.


Tweet of the Week

We scour Twitter each week to find an interesting/beautiful dashboard and show it off! #monitoringLove

YIKES! Glad it’s not – there’s good attention and bad attention.


Grafana Labs is Hiring!

We are passionate about open source software and thrive on tackling complex challenges to build the future. We ship code from every corner of the globe and love working with the community. If this sounds exciting, you’re in luck – WE’RE HIRING!

Check out our Open Positions


How are we doing?

Let us know if you’re finding these weekly roundups valuable. Submit a comment on this article below, or post something at our community forum. Find an article I haven’t included? Send it my way. Help us make timeShift better!

Follow us on Twitter, like us on Facebook, and join the Grafana Labs community.

Implementing Canary Deployments of AWS Lambda Functions with Alias Traffic Shifting

Post Syndicated from Chris Munns original https://aws.amazon.com/blogs/compute/implementing-canary-deployments-of-aws-lambda-functions-with-alias-traffic-shifting/

This post courtesy of Ryan Green, Software Development Engineer, AWS Serverless

The concepts of blue/green and canary deployments have been around for a while now and have been well-established as best-practices for reducing the risk of software deployments.

In a traditional, horizontally scaled application, copies of the application code are deployed to multiple nodes (instances, containers, on-premises servers, etc.), typically behind a load balancer. In these applications, deploying new versions of software to too many nodes at the same time can impact application availability as there may not be enough healthy nodes to service requests during the deployment. This aggressive approach to deployments also drastically increases the blast radius of software bugs introduced in the new version and does not typically give adequate time to safely assess the quality of the new version against production traffic.

In such applications, one commonly accepted solution to these problems is to slowly and incrementally roll out application software across the nodes in the fleet while simultaneously verifying application health (canary deployments). Another solution is to stand up an entirely different fleet and weight (or flip) traffic over to the new fleet after verification, ideally with some production traffic (blue/green). Some teams deploy to a single host (“one box environment”), where the new release can bake for some time before promotion to the rest of the fleet. Techniques like this enable the maintainers of complex systems to safely test in production while minimizing customer impact.

Enter Serverless

There is somewhat of an impedance mismatch when mapping these concepts to a serverless world. You can’t incrementally deploy your software across a fleet of servers when there are no servers!* In fact, even the term “deployment” takes on a different meaning with functions as a service (FaaS). In AWS Lambda, a “deployment” can be roughly modeled as a call to CreateFunction, UpdateFunctionCode, or UpdateAlias (I won’t get into the semantics of whether updating configuration counts as a deployment), all of which may affect the version of code that is invoked by clients.

The abstractions provided by Lambda remove the need for developers to be concerned about servers and Availability Zones, and this provides a powerful opportunity to greatly simplify the process of deploying software.
*Of course there are servers, but they are abstracted away from the developer.

Traffic shifting with Lambda aliases

Before the release of traffic shifting for Lambda aliases, deployments of a Lambda function could only be performed in a single “flip” by updating function code for version $LATEST, or by updating an alias to target a different function version. After the update propagates, typically within a few seconds, 100% of function invocations execute the new version. Implementing canary deployments with this model required the development of an additional routing layer, further adding development time, complexity, and invocation latency.
While rolling back a bad deployment of a Lambda function is a trivial operation and takes effect near instantaneously, deployments of new versions for critical functions can still be a potentially nerve-racking experience.

With the introduction of alias traffic shifting, it is now possible to trivially implement canary deployments of Lambda functions. By updating additional version weights on an alias, invocation traffic is routed to the new function versions based on the weight specified. Detailed CloudWatch metrics for the alias and version can be analyzed during the deployment, or other health checks performed, to ensure that the new version is healthy before proceeding.

Note: Sometimes the term “canary deployments” refers to the release of software to a subset of users. In the case of alias traffic shifting, the new version is released to some percentage of all users. It’s not possible to shard based on identity without adding an additional routing layer.

Examples

The simplest possible use of a canary deployment looks like the following:

# Update $LATEST version of function
aws lambda update-function-code --function-name myfunction ….

# Publish new version of function
aws lambda publish-version --function-name myfunction

# Point alias to new version, weighted at 5% (original version at 95% of traffic)
aws lambda update-alias --function-name myfunction --name myalias --routing-config '{"AdditionalVersionWeights" : {"2" : 0.05} }'

# Verify that the new version is healthy
…
# Set the primary version on the alias to the new version and reset the additional versions (100% weighted)
aws lambda update-alias --function-name myfunction --name myalias --function-version 2 --routing-config '{}'

This is begging to be automated! Here are a few options.

Simple deployment automation

This simple Python script runs as a Lambda function and deploys another function (how meta!) by incrementally increasing the weight of the new function version over a prescribed number of steps, while checking the health of the new version. If the health check fails, the alias is rolled back to its initial version. The health check is implemented as a simple check against the existence of Errors metrics in CloudWatch for the alias and new version.

GitHub aws-lambda-deploy repo

Install:

git clone https://github.com/awslabs/aws-lambda-deploy
cd aws-lambda-deploy
export BUCKET_NAME=[YOUR_S3_BUCKET_NAME_FOR_BUILD_ARTIFACTS]
./install.sh

Run:

# Rollout version 2 incrementally over 10 steps, with 120s between each step
aws lambda invoke --function-name SimpleDeployFunction --log-type Tail --payload \
  '{"function-name": "MyFunction",
  "alias-name": "MyAlias",
  "new-version": "2",
  "steps": 10,
  "interval" : 120,
  "type": "linear"
  }' output

Description of input parameters

  • function-name: The name of the Lambda function to deploy
  • alias-name: The name of the alias used to invoke the Lambda function
  • new-version: The version identifier for the new version to deploy
  • steps: The number of times the new version weight is increased
  • interval: The amount of time (in seconds) to wait between weight updates
  • type: The function to use to generate the weights. Supported values: “linear”

Because this runs as a Lambda function, it is subject to the maximum timeout of 5 minutes. This may be acceptable for many use cases, but to achieve a slower rollout of the new version, a different solution is required.

Step Functions workflow

This state machine performs essentially the same task as the simple deployment function, but it runs as an asynchronous workflow in AWS Step Functions. A nice property of Step Functions is that the maximum deployment timeout has now increased from 5 minutes to 1 year!

The step function incrementally updates the new version weight based on the steps parameter, waiting for some time based on the interval parameter, and performing health checks between updates. If the health check fails, the alias is rolled back to the original version and the workflow fails.

For example, to execute the workflow:

export STATE_MACHINE_ARN=`aws cloudformation describe-stack-resources --stack-name aws-lambda-deploy-stack --logical-resource-id DeployStateMachine --output text | cut  -d$'\t' -f3`

aws stepfunctions start-execution --state-machine-arn $STATE_MACHINE_ARN --input '{
  "function-name": "MyFunction",
  "alias-name": "MyAlias",
  "new-version": "2",
  "steps": 10,
  "interval": 120,
  "type": "linear"}'

Getting feedback on the deployment

Because the state machine runs asynchronously, retrieving feedback on the deployment requires polling for the execution status using DescribeExecution or implementing an asynchronous notification (using SNS or email, for example) from the Rollback or Finalize functions. A CloudWatch alarm could also be created to alarm based on the “ExecutionsFailed” metric for the state machine.

A note on health checks and observability

Weighted rollouts like this are considerably more successful if the code is being exercised and monitored continuously. In this example, it would help to have some automation continuously invoking the alias and reporting metrics on these invocations, such as client-side success rates and latencies.

The absence of Lambda Errors metrics used in these examples can be misleading if the function is not getting invoked. It’s also recommended to instrument your Lambda functions with custom metrics, in addition to Lambda’s built-in metrics, that can be used to monitor health during deployments.

Extensibility

These examples could be easily extended in various ways to support different use cases. For example:

  • Health check implementations: CloudWatch alarms, automatic invocations with payload assertions, querying external systems, etc.
  • Weight increase functions: Exponential, geometric progression, single canary step, etc.
  • Custom success/failure notifications: SNS, email, CI/CD systems, service discovery systems, etc.

Traffic shifting with SAM and CodeDeploy

Using the Lambda UpdateAlias operation with additional version weights provides a powerful primitive for you to implement custom traffic shifting solutions for Lambda functions.

For those not interested in building custom deployment solutions, AWS CodeDeploy provides an intuitive turn-key implementation of this functionality integrated directly into the Serverless Application Model. Traffic-shifted deployments can be declared in a SAM template, and CodeDeploy manages the function rollout as part of the CloudFormation stack update. CloudWatch alarms can also be configured to trigger a stack rollback if something goes wrong.

i.e.

MyFunction:
  Type: AWS::Serverless::Function
  Properties:
    FunctionName: MyFunction
    AutoPublishAlias: MyFunctionInvokeAlias
    DeploymentPreference:
      Type: Linear10PercentEvery1Minute
      Role:
        Fn::GetAtt: [ DeploymentRole, Arn ]
      Alarms:
       - { Ref: MyFunctionErrorsAlarm }
...

For more information about using CodeDeploy with SAM, see Automating Updates to Serverless Apps.

Conclusion

It is often the simple features that provide the most value. As I demonstrated in this post, serverless architectures allow the complex deployment orchestration used in traditional applications to be replaced with a simple Lambda function or Step Functions workflow. By allowing invocation traffic to be easily weighted to multiple function versions, Lambda alias traffic shifting provides a simple but powerful feature that I hope empowers you to easily implement safe deployment workflows for your Lambda functions.

AWS Fargate: A Product Overview

Post Syndicated from Deepak Dayama original https://aws.amazon.com/blogs/compute/aws-fargate-a-product-overview/

It was just about three years ago that AWS announced Amazon Elastic Container Service (Amazon ECS), to run and manage containers at scale on AWS. With Amazon ECS, you’ve been able to run your workloads at high scale and availability without having to worry about running your own cluster management and container orchestration software.

Today, AWS announced the availability of AWS Fargate – a technology that enables you to use containers as a fundamental compute primitive without having to manage the underlying instances. With Fargate, you don’t need to provision, configure, or scale virtual machines in your clusters to run containers. Fargate can be used with Amazon ECS today, with plans to support Amazon Elastic Container Service for Kubernetes (Amazon EKS) in the future.

Fargate has flexible configuration options so you can closely match your application needs and granular, per-second billing.

Amazon ECS with Fargate

Amazon ECS enables you to run containers at scale. This service also provides native integration into the AWS platform with VPC networking, load balancing, IAM, Amazon CloudWatch Logs, and CloudWatch metrics. These deep integrations make the Amazon ECS task a first-class object within the AWS platform.

To run tasks, you first need to stand up a cluster of instances, which involves picking the right types of instances and sizes, setting up Auto Scaling, and right-sizing the cluster for performance. With Fargate, you can leave all that behind and focus on defining your application and policies around permissions and scaling.

The same container management capabilities remain available so you can continue to scale your container deployments. With Fargate, the only entity to manage is the task. You don’t need to manage the instances or supporting software like Docker daemon or the Amazon ECS agent.

Fargate capabilities are available natively within Amazon ECS. This means that you don’t need to learn new API actions or primitives to run containers on Fargate.

Using Amazon ECS, Fargate is a launch type option. You continue to define the applications the same way by using task definitions. In contrast, the EC2 launch type gives you more control of your server clusters and provides a broader range of customization options.

For example, a RunTask command example is pasted below with the Fargate launch type:

ecs run-task --launch-type FARGATE --cluster fargate-test --task-definition nginx --network-configuration
"awsvpcConfiguration={subnets=[subnet-b563fcd3]}"

Key features of Fargate

Resource-based pricing and per second billing
You pay by the task size and only for the time for which resources are consumed by the task. The price for CPU and memory is charged on a per-second basis. There is a one-minute minimum charge.

Flexible configurations options
Fargate is available with 50 different combinations of CPU and memory to closely match your application needs. You can use 2 GB per vCPU anywhere up to 8 GB per vCPU for various configurations. Match your workload requirements closely, whether they are general purpose, compute, or memory optimized.

Networking
All Fargate tasks run within your own VPC. Fargate supports the recently launched awsvpc networking mode and the elastic network interface for a task is visible in the subnet where the task is running. This provides the separation of responsibility so you retain full control of networking policies for your applications via VPC features like security groups, routing rules, and NACLs. Fargate also supports public IP addresses.

Load Balancing
ECS Service Load Balancing  for the Application Load Balancer and Network Load Balancer is supported. For the Fargate launch type, you specify the IP addresses of the Fargate tasks to register with the load balancers.

Permission tiers
Even though there are no instances to manage with Fargate, you continue to group tasks into logical clusters. This allows you to manage who can run or view services within the cluster. The task IAM role is still applicable. Additionally, there is a new Task Execution Role that grants Amazon ECS permissions to perform operations such as pushing logs to CloudWatch Logs or pulling image from Amazon Elastic Container Registry (Amazon ECR).

Container Registry Support
Fargate provides seamless authentication to help pull images from Amazon ECR via the Task Execution Role. Similarly, if you are using a public repository like DockerHub, you can continue to do so.

Amazon ECS CLI
The Amazon ECS CLI provides high-level commands to help simplify to create and run Amazon ECS clusters, tasks, and services. The latest version of the CLI now supports running tasks and services with Fargate.

EC2 and Fargate Launch Type Compatibility
All Amazon ECS clusters are heterogeneous – you can run both Fargate and Amazon ECS tasks in the same cluster. This enables teams working on different applications to choose their own cadence of moving to Fargate, or to select a launch type that meets their requirements without breaking the existing model. You can make an existing ECS task definition compatible with the Fargate launch type and run it as a Fargate service, and vice versa. Choosing a launch type is not a one-way door!

Logging and Visibility
With Fargate, you can send the application logs to CloudWatch logs. Service metrics (CPU and Memory utilization) are available as part of CloudWatch metrics. AWS partners for visibility, monitoring and application performance management including Datadog, Aquasec, Splunk, Twistlock, and New Relic also support Fargate tasks.

Conclusion

Fargate enables you to run containers without having to manage the underlying infrastructure. Today, Fargate is availabe for Amazon ECS, and in 2018, Amazon EKS. Visit the Fargate product page to learn more, or get started in the AWS Console.

–Deepak Dayama

Object models

Post Syndicated from Eevee original https://eev.ee/blog/2017/11/28/object-models/

Anonymous asks, with dollars:

More about programming languages!

Well then!

I’ve written before about what I think objects are: state and behavior, which in practice mostly means method calls.

I suspect that the popular impression of what objects are, and also how they should work, comes from whatever C++ and Java happen to do. From that point of view, the whole post above is probably nonsense. If the baseline notion of “object” is a rigid definition woven tightly into the design of two massively popular languages, then it doesn’t even make sense to talk about what “object” should mean — it does mean the features of those languages, and cannot possibly mean anything else.

I think that’s a shame! It piles a lot of baggage onto a fairly simple idea. Polymorphism, for example, has nothing to do with objects — it’s an escape hatch for static type systems. Inheritance isn’t the only way to reuse code between objects, but it’s the easiest and fastest one, so it’s what we get. Frankly, it’s much closer to a speed tradeoff than a fundamental part of the concept.

We could do with more experimentation around how objects work, but that’s impossible in the languages most commonly thought of as object-oriented.

Here, then, is a (very) brief run through the inner workings of objects in four very dynamic languages. I don’t think I really appreciated objects until I’d spent some time with Python, and I hope this can help someone else whet their own appetite.

Python 3

Of the four languages I’m going to touch on, Python will look the most familiar to the Java and C++ crowd. For starters, it actually has a class construct.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
class Vector:
    def __init__(self, x, y):
        self.x = x
        self.y = y

    def __neg__(self):
        return Vector(-self.x, -self.y)

    def __div__(self, denom):
        return Vector(self.x / denom, self.y / denom)

    @property
    def magnitude(self):
        return (self.x ** 2 + self.y ** 2) ** 0.5

    def normalized(self):
        return self / self.magnitude

The __init__ method is an initializer, which is like a constructor but named differently (because the object already exists in a usable form by the time the initializer is called). Operator overloading is done by implementing methods with other special __dunder__ names. Properties can be created with @property, where the @ is syntax for applying a wrapper function to a function as it’s defined. You can do inheritance, even multiply:

1
2
3
4
class Foo(A, B, C):
    def bar(self, x, y, z):
        # do some stuff
        super().bar(x, y, z)

Cool, a very traditional object model.

Except… for some details.

Some details

For one, Python objects don’t have a fixed layout. Code both inside and outside the class can add or remove whatever attributes they want from whatever object they want. The underlying storage is just a dict, Python’s mapping type. (Or, rather, something like one. Also, it’s possible to change, which will probably be the case for everything I say here.)

If you create some attributes at the class level, you’ll start to get a peek behind the curtains:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
class Foo:
    values = []

    def add_value(self, value):
        self.values.append(value)

a = Foo()
b = Foo()
a.add_value('a')
print(a.values)  # ['a']
b.add_value('b')
print(b.values)  # ['a', 'b']

The [] assigned to values isn’t a default assigned to each object. In fact, the individual objects don’t know about it at all! You can use vars(a) to get at the underlying storage dict, and you won’t see a values entry in there anywhere.

Instead, values lives on the class, which is a value (and thus an object) in its own right. When Python is asked for self.values, it checks to see if self has a values attribute; in this case, it doesn’t, so Python keeps going and asks the class for one.

Python’s object model is secretly prototypical — a class acts as a prototype, as a shared set of fallback values, for its objects.

In fact, this is also how method calls work! They aren’t syntactically special at all, which you can see by separating the attribute lookup from the call.

1
2
3
print("abc".startswith("a"))  # True
meth = "abc".startswith
print(meth("a"))  # True

Reading obj.method looks for a method attribute; if there isn’t one on obj, Python checks the class. Here, it finds one: it’s a function from the class body.

Ah, but wait! In the code I just showed, meth seems to “know” the object it came from, so it can’t just be a plain function. If you inspect the resulting value, it claims to be a “bound method” or “built-in method” rather than a function, too. Something funny is going on here, and that funny something is the descriptor protocol.

Descriptors

Python allows attributes to implement their own custom behavior when read from or written to. Such an attribute is called a descriptor. I’ve written about them before, but here’s a quick overview.

If Python looks up an attribute, finds it in a class, and the value it gets has a __get__ method… then instead of using that value, Python will use the return value of its __get__ method.

The @property decorator works this way. The magnitude property in my original example was shorthand for doing this:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
class MagnitudeDescriptor:
    def __get__(self, instance, owner):
        if instance is None:
            return self
        return (instance.x ** 2 + instance.y ** 2) ** 0.5

class Vector:
    def __init__(self, x, y):
        self.x = x
        self.y = y

    magnitude = MagnitudeDescriptor()

When you ask for somevec.magnitude, Python checks somevec but doesn’t find magnitude, so it consults the class instead. The class does have a magnitude, and it’s a value with a __get__ method, so Python calls that method and somevec.magnitude evaluates to its return value. (The instance is None check is because __get__ is called even if you get the descriptor directly from the class via Vector.magnitude. A descriptor intended to work on instances can’t do anything useful in that case, so the convention is to return the descriptor itself.)

You can also intercept attempts to write to or delete an attribute, and do absolutely whatever you want instead. But note that, similar to operating overloading in Python, the descriptor must be on a class; you can’t just slap one on an arbitrary object and have it work.

This brings me right around to how “bound methods” actually work. Functions are descriptors! The function type implements __get__, and when a function is retrieved from a class via an instance, that __get__ bundles the function and the instance together into a tiny bound method object. It’s essentially:

1
2
3
4
5
class FunctionType:
    def __get__(self, instance, owner):
        if instance is None:
            return self
        return functools.partial(self, instance)

The self passed as the first argument to methods is not special or magical in any way. It’s built out of a few simple pieces that are also readily accessible to Python code.

Note also that because obj.method() is just an attribute lookup and a call, Python doesn’t actually care whether method is a method on the class or just some callable thing on the object. You won’t get the auto-self behavior if it’s on the object, but otherwise there’s no difference.

More attribute access, and the interesting part

Descriptors are one of several ways to customize attribute access. Classes can implement __getattr__ to intervene when an attribute isn’t found on an object; __setattr__ and __delattr__ to intervene when any attribute is set or deleted; and __getattribute__ to implement unconditional attribute access. (That last one is a fantastic way to create accidental recursion, since any attribute access you do within __getattribute__ will of course call __getattribute__ again.)

Here’s what I really love about Python. It might seem like a magical special case that descriptors only work on classes, but it really isn’t. You could implement exactly the same behavior yourself, in pure Python, using only the things I’ve just told you about. Classes are themselves objects, remember, and they are instances of type, so the reason descriptors only work on classes is that type effectively does this:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
class type:
    def __getattribute__(self, name):
        value = super().__getattribute__(name)
        # like all op overloads, __get__ must be on the type, not the instance
        ty = type(value)
        if hasattr(ty, '__get__'):
            # it's a descriptor!  this is a class access so there is no instance
            return ty.__get__(value, None, self)
        else:
            return value

You can even trivially prove to yourself that this is what’s going on by skipping over types behavior:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
class Descriptor:
    def __get__(self, instance, owner):
        print('called!')

class Foo:
    bar = Descriptor()

Foo.bar  # called!
type.__getattribute__(Foo, 'bar')  # called!
object.__getattribute__(Foo, 'bar')  # ...

And that’s not all! The mysterious super function, used to exhaustively traverse superclass method calls even in the face of diamond inheritance, can also be expressed in pure Python using these primitives. You could write your own superclass calling convention and use it exactly the same way as super.

This is one of the things I really like about Python. Very little of it is truly magical; virtually everything about the object model exists in the types rather than the language, which means virtually everything can be customized in pure Python.

Class creation and metaclasses

A very brief word on all of this stuff, since I could talk forever about Python and I have three other languages to get to.

The class block itself is fairly interesting. It looks like this:

1
2
class Name(*bases, **kwargs):
    # code

I’ve said several times that classes are objects, and in fact the class block is one big pile of syntactic sugar for calling type(...) with some arguments to create a new type object.

The Python documentation has a remarkably detailed description of this process, but the gist is:

  • Python determines the type of the new class — the metaclass — by looking for a metaclass keyword argument. If there isn’t one, Python uses the “lowest” type among the provided base classes. (If you’re not doing anything special, that’ll just be type, since every class inherits from object and object is an instance of type.)

  • Python executes the class body. It gets its own local scope, and any assignments or method definitions go into that scope.

  • Python now calls type(name, bases, attrs, **kwargs). The name is whatever was right after class; the bases are position arguments; and attrs is the class body’s local scope. (This is how methods and other class attributes end up on the class.) The brand new type is then assigned to Name.

Of course, you can mess with most of this. You can implement __prepare__ on a metaclass, for example, to use a custom mapping as storage for the local scope — including any reads, which allows for some interesting shenanigans. The only part you can’t really implement in pure Python is the scoping bit, which has a couple extra rules that make sense for classes. (In particular, functions defined within a class block don’t close over the class body; that would be nonsense.)

Object creation

Finally, there’s what actually happens when you create an object — including a class, which remember is just an invocation of type(...).

Calling Foo(...) is implemented as, well, a call. Any type can implement calls with the __call__ special method, and you’ll find that type itself does so. It looks something like this:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
# oh, a fun wrinkle that's hard to express in pure python: type is a class, so
# it's an instance of itself
class type:
    def __call__(self, *args, **kwargs):
        # remember, here 'self' is a CLASS, an instance of type.
        # __new__ is a true constructor: object.__new__ allocates storage
        # for a new blank object
        instance = self.__new__(self, *args, **kwargs)
        # you can return whatever you want from __new__ (!), and __init__
        # is only called on it if it's of the right type
        if isinstance(instance, self):
            instance.__init__(*args, **kwargs)
        return instance

Again, you can trivially confirm this by asking any type for its __call__ method. Assuming that type doesn’t implement __call__ itself, you’ll get back a bound version of types implementation.

1
2
>>> list.__call__
<method-wrapper '__call__' of type object at 0x7fafb831a400>

You can thus implement __call__ in your own metaclass to completely change how subclasses are created — including skipping the creation altogether, if you like.

And… there’s a bunch of stuff I haven’t even touched on.

The Python philosophy

Python offers something that, on the surface, looks like a “traditional” class/object model. Under the hood, it acts more like a prototypical system, where failed attribute lookups simply defer to a superclass or metaclass.

The language also goes to almost superhuman lengths to expose all of its moving parts. Even the prototypical behavior is an implementation of __getattribute__ somewhere, which you are free to completely replace in your own types. Proxying and delegation are easy.

Also very nice is that these features “bundle” well, by which I mean a library author can do all manner of convoluted hijinks, and a consumer of that library doesn’t have to see any of it or understand how it works. You only need to inherit from a particular class (which has a metaclass), or use some descriptor as a decorator, or even learn any new syntax.

This meshes well with Python culture, which is pretty big on the principle of least surprise. These super-advanced features tend to be tightly confined to single simple features (like “makes a weak attribute“) or cordoned with DSLs (e.g., defining a form/struct/database table with a class body). In particular, I’ve never seen a metaclass in the wild implement its own __call__.

I have mixed feelings about that. It’s probably a good thing overall that the Python world shows such restraint, but I wonder if there are some very interesting possibilities we’re missing out on. I implemented a metaclass __call__ myself, just once, in an entity/component system that strove to minimize fuss when communicating between components. It never saw the light of day, but I enjoyed seeing some new things Python could do with the same relatively simple syntax. I wouldn’t mind seeing, say, an object model based on composition (with no inheritance) built atop Python’s primitives.

Lua

Lua doesn’t have an object model. Instead, it gives you a handful of very small primitives for building your own object model. This is pretty typical of Lua — it’s a very powerful language, but has been carefully constructed to be very small at the same time. I’ve never encountered anything else quite like it, and “but it starts indexing at 1!” really doesn’t do it justice.

The best way to demonstrate how objects work in Lua is to build some from scratch. We need two key features. The first is metatables, which bear a passing resemblance to Python’s metaclasses.

Tables and metatables

The table is Lua’s mapping type and its primary data structure. Keys can be any value other than nil. Lists are implemented as tables whose keys are consecutive integers starting from 1. Nothing terribly surprising. The dot operator is sugar for indexing with a string key.

1
2
3
4
5
local t = { a = 1, b = 2 }
print(t['a'])  -- 1
print(t.b)  -- 2
t.c = 3
print(t['c'])  -- 3

A metatable is a table that can be associated with another value (usually another table) to change its behavior. For example, operator overloading is implemented by assigning a function to a special key in a metatable.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
local t = { a = 1, b = 2 }
--print(t + 0)  -- error: attempt to perform arithmetic on a table value

local mt = {
    __add = function(left, right)
        return 12
    end,
}
setmetatable(t, mt)
print(t + 0)  -- 12

Now, the interesting part: one of the special keys is __index, which is consulted when the base table is indexed by a key it doesn’t contain. Here’s a table that claims every key maps to itself.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
local t = {}
local mt = {
    __index = function(table, key)
        return key
    end,
}
setmetatable(t, mt)
print(t.foo)  -- foo
print(t.bar)  -- bar
print(t[3])  -- 3

__index doesn’t have to be a function, either. It can be yet another table, in which case that table is simply indexed with the key. If the key still doesn’t exist and that table has a metatable with an __index, the process repeats.

With this, it’s easy to have several unrelated tables that act as a single table. Call the base table an object, fill the __index table with functions and call it a class, and you have half of an object system. You can even get prototypical inheritance by chaining __indexes together.

At this point things are a little confusing, since we have at least three tables going on, so here’s a diagram. Keep in mind that Lua doesn’t actually have anything called an “object”, “class”, or “method” — those are just convenient nicknames for a particular structure we might build with Lua’s primitives.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
                    ╔═══════════╗        ...
                    ║ metatable ║         ║
                    ╟───────────╢   ┌─────╨───────────────────────┐
                    ║ __index   ╫───┤ lookup table ("superclass") │
                    ╚═══╦═══════╝   ├─────────────────────────────┤
  ╔═══════════╗         ║           │ some other method           ┼─── function() ... end
  ║ metatable ║         ║           └─────────────────────────────┘
  ╟───────────╢   ┌─────╨──────────────────┐
  ║ __index   ╫───┤ lookup table ("class") │
  ╚═══╦═══════╝   ├────────────────────────┤
      ║           │ some method            ┼─── function() ... end
      ║           └────────────────────────┘
┌─────╨─────────────────┐
│ base table ("object") │
└───────────────────────┘

Note that a metatable is not the same as a class; it defines behavior, not methods. Conversely, if you try to use a class directly as a metatable, it will probably not do much. (This is pretty different from e.g. Python, where operator overloads are just methods with funny names. One nice thing about the Lua approach is that you can keep interface-like functionality separate from methods, and avoid clogging up arbitrary objects’ namespaces. You could even use a dummy table as a key and completely avoid name collisions.)

Anyway, code!

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
local class = {
    foo = function(a)
        print("foo got", a)
    end,
}
local mt = { __index = class }
-- setmetatable returns its first argument, so this is nice shorthand
local obj1 = setmetatable({}, mt)
local obj2 = setmetatable({}, mt)
obj1.foo(7)  -- foo got 7
obj2.foo(9)  -- foo got 9

Wait, wait, hang on. Didn’t I call these methods? How do they get at the object? Maybe Lua has a magical this variable?

Methods, sort of

Not quite, but this is where the other key feature comes in: method-call syntax. It’s the lightest touch of sugar, just enough to have method invocation.

1
2
3
4
5
6
7
8
9
-- note the colon!
a:b(c, d, ...)

-- exactly equivalent to this
-- (except that `a` is only evaluated once)
a.b(a, c, d, ...)

-- which of course is really this
a["b"](a, c, d, ...)

Now we can write methods that actually do something.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
local class = {
    bar = function(self)
        print("our score is", self.score)
    end,
}
local mt = { __index = class }
local obj1 = setmetatable({ score = 13 }, mt)
local obj2 = setmetatable({ score = 25 }, mt)
obj1:bar()  -- our score is 13
obj2:bar()  -- our score is 25

And that’s all you need. Much like Python, methods and data live in the same namespace, and Lua doesn’t care whether obj:method() finds a function on obj or gets one from the metatable’s __index. Unlike Python, the function will be passed self either way, because self comes from the use of : rather than from the lookup behavior.

(Aside: strictly speaking, any Lua value can have a metatable — and if you try to index a non-table, Lua will always consult the metatable’s __index. Strings all have the string library as a metatable, so you can call methods on them: try ("%s %s"):format(1, 2). I don’t think Lua lets user code set the metatable for non-tables, so this isn’t that interesting, but if you’re writing Lua bindings from C then you can wrap your pointers in metatables to give them methods implemented in C.)

Bringing it all together

Of course, writing all this stuff every time is a little tedious and error-prone, so instead you might want to wrap it all up inside a little function. No problem.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
local function make_object(body)
    -- create a metatable
    local mt = { __index = body }
    -- create a base table to serve as the object itself
    local obj = setmetatable({}, mt)
    -- and, done
    return obj
end

-- you can leave off parens if you're only passing in 
local Dog = {
    -- this acts as a "default" value; if obj.barks is missing, __index will
    -- kick in and find this value on the class.  but if obj.barks is assigned
    -- to, it'll go in the object and shadow the value here.
    barks = 0,

    bark = function(self)
        self.barks = self.barks + 1
        print("woof!")
    end,
}

local mydog = make_object(Dog)
mydog:bark()  -- woof!
mydog:bark()  -- woof!
mydog:bark()  -- woof!
print(mydog.barks)  -- 3
print(Dog.barks)  -- 0

It works, but it’s fairly barebones. The nice thing is that you can extend it pretty much however you want. I won’t reproduce an entire serious object system here — lord knows there are enough of them floating around — but the implementation I have for my LÖVE games lets me do this:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
local Animal = Object:extend{
    cries = 0,
}

-- called automatically by Object
function Animal:init()
    print("whoops i couldn't think of anything interesting to put here")
end

-- this is just nice syntax for adding a first argument called 'self', then
-- assigning this function to Animal.cry
function Animal:cry()
    self.cries = self.cries + 1
end

local Cat = Animal:extend{}

function Cat:cry()
    print("meow!")
    Cat.__super.cry(self)
end

local cat = Cat()
cat:cry()  -- meow!
cat:cry()  -- meow!
print(cat.cries)  -- 2

When I say you can extend it however you want, I mean that. I could’ve implemented Python (2)-style super(Cat, self):cry() syntax; I just never got around to it. I could even make it work with multiple inheritance if I really wanted to — or I could go the complete opposite direction and only implement composition. I could implement descriptors, customizing the behavior of individual table keys. I could add pretty decent syntax for composition/proxying. I am trying very hard to end this section now.

The Lua philosophy

Lua’s philosophy is to… not have a philosophy? It gives you the bare minimum to make objects work, and you can do absolutely whatever you want from there. Lua does have something resembling prototypical inheritance, but it’s not so much a first-class feature as an emergent property of some very simple tools. And since you can make __index be a function, you could avoid the prototypical behavior and do something different entirely.

The very severe downside, of course, is that you have to find or build your own object system — which can get pretty confusing very quickly, what with the multiple small moving parts. Third-party code may also have its own object system with subtly different behavior. (Though, in my experience, third-party code tries very hard to avoid needing an object system at all.)

It’s hard to say what the Lua “culture” is like, since Lua is an embedded language that’s often a little different in each environment. I imagine it has a thousand millicultures, instead. I can say that the tedium of building my own object model has led me into something very “traditional”, with prototypical inheritance and whatnot. It’s partly what I’m used to, but it’s also just really dang easy to get working.

Likewise, while I love properties in Python and use them all the dang time, I’ve yet to use a single one in Lua. They wouldn’t be particularly hard to add to my object model, but having to add them myself (or shop around for an object model with them and also port all my code to use it) adds a huge amount of friction. I’ve thought about designing an interesting ECS with custom object behavior, too, but… is it really worth the effort? For all the power and flexibility Lua offers, the cost is that by the time I have something working at all, I’m too exhausted to actually use any of it.

JavaScript

JavaScript is notable for being preposterously heavily used, yet not having a class block.

Well. Okay. Yes. It has one now. It didn’t for a very long time, and even the one it has now is sugar.

Here’s a vector class again:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
class Vector {
    constructor(x, y) {
        this.x = x;
        this.y = y;
    }

    get magnitude() {
        return Math.sqrt(this.x * this.x + this.y * this.y);
    }

    dot(other) {
        return this.x * other.x + this.y * other.y;
    }
}

In “classic” JavaScript, this would be written as:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
function Vector(x, y) {
    this.x = x;
    this.y = y;
}

Object.defineProperty(Vector.prototype, 'magnitude', {
    configurable: true,
    enumerable: true,
    get: function() {
        return Math.sqrt(this.x * this.x + this.y * this.y);
    },
});


Vector.prototype.dot = function(other) {
    return this.x * other.x + this.y * other.y;
};

Hm, yes. I can see why they added class.

The JavaScript model

In JavaScript, a new type is defined in terms of a function, which is its constructor.

Right away we get into trouble here. There is a very big difference between these two invocations, which I actually completely forgot about just now after spending four hours writing about Python and Lua:

1
2
let vec = Vector(3, 4);
let vec = new Vector(3, 4);

The first calls the function Vector. It assigns some properties to this, which here is going to be window, so now you have a global x and y. It then returns nothing, so vec is undefined.

The second calls Vector with this set to a new empty object, then evaluates to that object. The result is what you’d actually expect.

(You can detect this situation with the strange new.target expression, but I have never once remembered to do so.)

From here, we have true, honest-to-god, first-class prototypical inheritance. The word “prototype” is even right there. When you write this:

1
vec.dot(vec2)

JavaScript will look for dot on vec and (presumably) not find it. It then consults vecs prototype, an object you can see for yourself by using Object.getPrototypeOf(). Since vec is a Vector, its prototype is Vector.prototype.

I stress that Vector.prototype is not the prototype for Vector. It’s the prototype for instances of Vector.

(I say “instance”, but the true type of vec here is still just object. If you want to find Vector, it’s automatically assigned to the constructor property of its own prototype, so it’s available as vec.constructor.)

Of course, Vector.prototype can itself have a prototype, in which case the process would continue if dot were not found. A common (and, arguably, very bad) way to simulate single inheritance is to set Class.prototype to an instance of a superclass to get the prototype right, then tack on the methods for Class. Nowadays we can do Object.create(Superclass.prototype).

Now that I’ve been through Python and Lua, though, this isn’t particularly surprising. I kinda spoiled it.

I suppose one difference in JavaScript is that you can tack arbitrary attributes directly onto Vector all you like, and they will remain invisible to instances since they aren’t in the prototype chain. This is kind of backwards from Lua, where you can squirrel stuff away in the metatable.

Another difference is that every single object in JavaScript has a bunch of properties already tacked on — the ones in Object.prototype. Every object (and by “object” I mean any mapping) has a prototype, and that prototype defaults to Object.prototype, and it has a bunch of ancient junk like isPrototypeOf.

(Nit: it’s possible to explicitly create an object with no prototype via Object.create(null).)

Like Lua, and unlike Python, JavaScript doesn’t distinguish between keys found on an object and keys found via a prototype. Properties can be defined on prototypes with Object.defineProperty(), but that works just as well directly on an object, too. JavaScript doesn’t have a lot of operator overloading, but some things like Symbol.iterator also work on both objects and prototypes.

About this

You may, at this point, be wondering what this is. Unlike Lua and Python (and the last language below), this is a special built-in value — a context value, invisibly passed for every function call.

It’s determined by where the function came from. If the function was the result of an attribute lookup, then this is set to the object containing that attribute. Otherwise, this is set to the global object, window. (You can also set this to whatever you want via the call method on functions.)

This decision is made lexically, i.e. from the literal source code as written. There are no Python-style bound methods. In other words:

1
2
3
4
5
// this = obj
obj.method()
// this = window
let meth = obj.method
meth()

Also, because this is reassigned on every function call, it cannot be meaningfully closed over, which makes using closures within methods incredibly annoying. The old approach was to assign this to some other regular name like self (which got syntax highlighting since it’s also a built-in name in browsers); then we got Function.bind, which produced a callable thing with a fixed context value, which was kind of nice; and now finally we have arrow functions, which explicitly close over the current this when they’re defined and don’t change it when called. Phew.

Class syntax

I already showed class syntax, and it’s really just one big macro for doing all the prototype stuff The Right Way. It even prevents you from calling the type without new. The underlying model is exactly the same, and you can inspect all the parts.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
class Vector { ... }

console.log(Vector.prototype);  // { dot: ..., magnitude: ..., ... }
let vec = new Vector(3, 4);
console.log(Object.getPrototypeOf(vec));  // same as Vector.prototype

// i don't know why you would subclass vector but let's roll with it
class Vectest extends Vector { ... }

console.log(Vectest.prototype);  // { ... }
console.log(Object.getPrototypeOf(Vectest.prototype))  // same as Vector.prototype

Alas, class syntax has a couple shortcomings. You can’t use the class block to assign arbitrary data to either the type object or the prototype — apparently it was deemed too confusing that mutations would be shared among instances. Which… is… how prototypes work. How Python works. How JavaScript itself, one of the most popular languages of all time, has worked for twenty-two years. Argh.

You can still do whatever assignment you want outside of the class block, of course. It’s just a little ugly, and not something I’d think to look for with a sugary class.

A more subtle result of this behavior is that a class block isn’t quite the same syntax as an object literal. The check for data isn’t a runtime thing; class Foo { x: 3 } fails to parse. So JavaScript now has two largely but not entirely identical styles of key/value block.

Attribute access

Here’s where things start to come apart at the seams, just a little bit.

JavaScript doesn’t really have an attribute protocol. Instead, it has two… extension points, I suppose.

One is Object.defineProperty, seen above. For common cases, there’s also the get syntax inside a property literal, which does the same thing. But unlike Python’s @property, these aren’t wrappers around some simple primitives; they are the primitives. JavaScript is the only language of these four to have “property that runs code on access” as a completely separate first-class concept.

If you want to intercept arbitrary attribute access (and some kinds of operators), there’s a completely different primitive: the Proxy type. It doesn’t let you intercept attribute access or operators; instead, it produces a wrapper object that supports interception and defers to the wrapped object by default.

It’s cool to see composition used in this way, but also, extremely weird. If you want to make your own type that overloads in or calling, you have to return a Proxy that wraps your own type, rather than actually returning your own type. And (unlike the other three languages in this post) you can’t return a different type from a constructor, so you have to throw that away and produce objects only from a factory. And instanceof would be broken, but you can at least fix that with Symbol.hasInstance — which is really operator overloading, implement yet another completely different way.

I know the design here is a result of legacy and speed — if any object could intercept all attribute access, then all attribute access would be slowed down everywhere. Fair enough. It still leaves the surface area of the language a bit… bumpy?

The JavaScript philosophy

It’s a little hard to tell. The original idea of prototypes was interesting, but it was hidden behind some very awkward syntax. Since then, we’ve gotten a bunch of extra features awkwardly bolted on to reflect the wildly varied things the built-in types and DOM API were already doing. We have class syntax, but it’s been explicitly designed to avoid exposing the prototype parts of the model.

I admit I don’t do a lot of heavy JavaScript, so I might just be overlooking it, but I’ve seen virtually no code that makes use of any of the recent advances in object capabilities. Forget about custom iterators or overloading call; I can’t remember seeing any JavaScript in the wild that even uses properties yet. I don’t know if everyone’s waiting for sufficient browser support, nobody knows about them, or nobody cares.

The model has advanced recently, but I suspect JavaScript is still shackled to its legacy of “something about prototypes, I don’t really get it, just copy the other code that’s there” as an object model. Alas! Prototypes are so good. Hopefully class syntax will make it a bit more accessible, as it has in Python.

Perl 5

Perl 5 also doesn’t have an object system and expects you to build your own. But where Lua gives you two simple, powerful tools for building one, Perl 5 feels more like a puzzle with half the pieces missing. Clearly they were going for something, but they only gave you half of it.

In brief, a Perl object is a reference that has been blessed with a package.

I need to explain a few things. Honestly, one of the biggest problems with the original Perl object setup was how many strange corners and unique jargon you had to understand just to get off the ground.

(If you want to try running any of this code, you should stick a use v5.26; as the first line. Perl is very big on backwards compatibility, so you need to opt into breaking changes, and even the mundane say builtin is behind a feature gate.)

References

A reference in Perl is sort of like a pointer, but its main use is very different. See, Perl has the strange property that its data structures try very hard to spill their contents all over the place. Despite having dedicated syntax for arrays — @foo is an array variable, distinct from the single scalar variable $foo — it’s actually impossible to nest arrays.

1
2
3
my @foo = (1, 2, 3, 4);
my @bar = (@foo, @foo);
# @bar is now a flat list of eight items: 1, 2, 3, 4, 1, 2, 3, 4

The idea, I guess, is that an array is not one thing. It’s not a container, which happens to hold multiple things; it is multiple things. Anywhere that expects a single value, such as an array element, cannot contain an array, because an array fundamentally is not a single value.

And so we have “references”, which are a form of indirection, but also have the nice property that they’re single values. They add containment around arrays, and in general they make working with most of Perl’s primitive types much more sensible. A reference to a variable can be taken with the \ operator, or you can use [ ... ] and { ... } to directly create references to anonymous arrays or hashes.

1
2
3
my @foo = (1, 2, 3, 4);
my @bar = (\@foo, \@foo);
# @bar is now a nested list of two items: [1, 2, 3, 4], [1, 2, 3, 4]

(Incidentally, this is the sole reason I initially abandoned Perl for Python. Non-trivial software kinda requires nesting a lot of data structures, so you end up with references everywhere, and the syntax for going back and forth between a reference and its contents is tedious and ugly.)

A Perl object must be a reference. Perl doesn’t care what kind of reference — it’s usually a hash reference, since hashes are a convenient place to store arbitrary properties, but it could just as well be a reference to an array, a scalar, or even a sub (i.e. function) or filehandle.

I’m getting a little ahead of myself. First, the other half: blessing and packages.

Packages and blessing

Perl packages are just namespaces. A package looks like this:

1
2
3
4
5
6
7
package Foo::Bar;

sub quux {
    say "hi from quux!";
}

# now Foo::Bar::quux() can be called from anywhere

Nothing shocking, right? It’s just a named container. A lot of the details are kind of weird, like how a package exists in some liminal quasi-value space, but the basic idea is a Bag Of Stuff.

The final piece is “blessing,” which is Perl’s funny name for binding a package to a reference. A very basic class might look like this:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
package Vector;

# the name 'new' is convention, not special
sub new {
    # perl argument passing is weird, don't ask
    my ($class, $x, $y) = @_;

    # create the object itself -- here, unusually, an array reference makes sense
    my $self = [ $x, $y ];

    # associate the package with that reference
    # note that $class here is just the regular string, 'Vector'
    bless $self, $class;

    return $self;
}

sub x {
    my ($self) = @_;
    return $self->[0];
}

sub y {
    my ($self) = @_;
    return $self->[1];
}

sub magnitude {
    my ($self) = @_;
    return sqrt($self->x ** 2 + $self->y ** 2);
}

# switch back to the "default" package
package main;

# -> is method call syntax, which passes the invocant as the first argument;
# for a package, that's just the package name
my $vec = Vector->new(3, 4);
say $vec->magnitude;  # 5

A few things of note here. First, $self->[0] has nothing to do with objects; it’s normal syntax for getting the value of a index 0 out of an array reference called $self. (Most classes are based on hashrefs and would use $self->{value} instead.) A blessed reference is still a reference and can be treated like one.

In general, -> is Perl’s dereferencey operator, but its exact behavior depends on what follows. If it’s followed by brackets, then it’ll apply the brackets to the thing in the reference: ->{} to index a hash reference, ->[] to index an array reference, and ->() to call a function reference.

But if -> is followed by an identifier, then it’s a method call. For packages, that means calling a function in the package and passing the package name as the first argument. For objects — blessed references — that means calling a function in the associated package and passing the object as the first argument.

This is a little weird! A blessed reference is a superposition of two things: its normal reference behavior, and some completely orthogonal object behavior. Also, object behavior has no notion of methods vs data; it only knows about methods. Perl lets you omit parentheses in a lot of places, including when calling a method with no arguments, so $vec->magnitude is really $vec->magnitude().

Perl’s blessing bears some similarities to Lua’s metatables, but ultimately Perl is much closer to Ruby’s “message passing” approach than the above three languages’ approaches of “get me something and maybe it’ll be callable”. (But this is no surprise — Ruby is a spiritual successor to Perl 5.)

All of this leads to one little wrinkle: how do you actually expose data? Above, I had to write x and y methods. Am I supposed to do that for every single attribute on my type?

Yes! But don’t worry, there are third-party modules to help with this incredibly fundamental task. Take Class::Accessor::Fast, so named because it’s faster than Class::Accessor:

1
2
3
package Foo;
use base qw(Class::Accessor::Fast);
__PACKAGE__->mk_accessors(qw(fred wilma barney));

(__PACKAGE__ is the lexical name of the current package; qw(...) is a list literal that splits its contents on whitespace.)

This assumes you’re using a hashref with keys of the same names as the attributes. $obj->fred will return the fred key from your hashref, and $obj->fred(4) will change it to 4.

You also, somewhat bizarrely, have to inherit from Class::Accessor::Fast. Speaking of which,

Inheritance

Inheritance is done by populating the package-global @ISA array with some number of (string) names of parent packages. Most code instead opts to write use base ...;, which does the same thing. Or, more commonly, use parent ...;, which… also… does the same thing.

Every package implicitly inherits from UNIVERSAL, which can be freely modified by Perl code.

A method can call its superclass method with the SUPER:: pseudo-package:

1
2
3
4
sub foo {
    my ($self) = @_;
    $self->SUPER::foo;
}

However, this does a depth-first search, which means it almost certainly does the wrong thing when faced with multiple inheritance. For a while the accepted solution involved a third-party module, but Perl eventually grew an alternative you have to opt into: C3, which may be more familiar to you as the order Python uses.

1
2
3
4
5
6
use mro 'c3';

sub foo {
    my ($self) = @_;
    $self->next::method;
}

Offhand, I’m not actually sure how next::method works, seeing as it was originally implemented in pure Perl code. I suspect it involves peeking at the caller’s stack frame. If so, then this is a very different style of customizability from e.g. Python — the MRO was never intended to be pluggable, and the use of a special pseudo-package means it isn’t really, but someone was determined enough to make it happen anyway.

Operator overloading and whatnot

Operator overloading looks a little weird, though really it’s pretty standard Perl.

1
2
3
4
5
6
7
8
package MyClass;

use overload '+' => \&_add;

sub _add {
    my ($self, $other, $swap) = @_;
    ...
}

use overload here is a pragma, where “pragma” means “regular-ass module that does some wizardry when imported”.

\&_add is how you get a reference to the _add sub so you can pass it to the overload module. If you just said &_add or _add, that would call it.

And that’s it; you just pass a map of operators to functions to this built-in module. No worry about name clashes or pollution, which is pretty nice. You don’t even have to give references to functions that live in the package, if you don’t want them to clog your namespace; you could put them in another package, or even inline them anonymously.

One especially interesting thing is that Perl lets you overload every operator. Perl has a lot of operators. It considers some math builtins like sqrt and trig functions to be operators, or at least operator-y enough that you can overload them. You can also overload the “file text” operators, such as -e $path to test whether a file exists. You can overload conversions, including implicit conversion to a regex. And most fascinating to me, you can overload dereferencing — that is, the thing Perl does when you say $hashref->{key} to get at the underlying hash. So a single object could pretend to be references of multiple different types, including a subref to implement callability. Neat.

Somewhat related: you can overload basic operators (indexing, etc.) on basic types (not references!) with the tie function, which is designed completely differently and looks for methods with fixed names. Go figure.

You can intercept calls to nonexistent methods by implementing a function called AUTOLOAD, within which the $AUTOLOAD global will contain the name of the method being called. Originally this feature was, I think, intended for loading binary components or large libraries on-the-fly only when needed, hence the name. Offhand I’m not sure I ever saw it used the way __getattr__ is used in Python.

Is there a way to intercept all method calls? I don’t think so, but it is Perl, so I must be forgetting something.

Actually no one does this any more

Like a decade ago, a council of elder sages sat down and put together a whole whizbang system that covers all of it: Moose.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
package Vector;
use Moose;

has x => (is => 'rw', isa => 'Int');
has y => (is => 'rw', isa => 'Int');

sub magnitude {
    my ($self) = @_;
    return sqrt($self->x ** 2 + $self->y ** 2);
}

Moose has its own way to do pretty much everything, and it’s all built on the same primitives. Moose also adds metaclasses, somehow, despite that the underlying model doesn’t actually support them? I’m not entirely sure how they managed that, but I do remember doing some class introspection with Moose and it was much nicer than the built-in way.

(If you’re wondering, the built-in way begins with looking at the hash called %Vector::. No, that’s not a typo.)

I really cannot stress enough just how much stuff Moose does, but I don’t want to delve into it here since Moose itself is not actually the language model.

The Perl philosophy

I hope you can see what I meant with what I first said about Perl, now. It has multiple inheritance with an MRO, but uses the wrong one by default. It has extensive operator overloading, which looks nothing like how inheritance works, and also some of it uses a totally different mechanism with special method names instead. It only understands methods, not data, leaving you to figure out accessors by hand.

There’s 70% of an object system here with a clear general design it was gunning for, but none of the pieces really look anything like each other. It’s weird, in a distinctly Perl way.

The result is certainly flexible, at least! It’s especially cool that you can use whatever kind of reference you want for storage, though even as I say that, I acknowledge it’s no different from simply subclassing list or something in Python. It feels different in Perl, but maybe only because it looks so different.

I haven’t written much Perl in a long time, so I don’t know what the community is like any more. Moose was already ubiquitous when I left, which you’d think would let me say “the community mostly focuses on the stuff Moose can do” — but even a decade ago, Moose could already do far more than I had ever seen done by hand in Perl. It’s always made a big deal out of roles (read: interfaces), for instance, despite that I’d never seen anyone care about them in Perl before Moose came along. Maybe their presence in Moose has made them more popular? Who knows.

Also, I wrote Perl seriously, but in the intervening years I’ve only encountered people who only ever used Perl for one-offs. Maybe it’ll come as a surprise to a lot of readers that Perl has an object model at all.

End

Well, that was fun! I hope any of that made sense.

Special mention goes to Rust, which doesn’t have an object model you can fiddle with at runtime, but does do things a little differently.

It’s been really interesting thinking about how tiny differences make a huge impact on what people do in practice. Take the choice of storage in Perl versus Python. Perl’s massively common URI class uses a string as the storage, nothing else; I haven’t seen anything like that in Python aside from markupsafe, which is specifically designed as a string type. I would guess this is partly because Perl makes you choose — using a hashref is an obvious default, but you have to make that choice one way or the other. In Python (especially 3), inheriting from object and getting dict-based storage is the obvious thing to do; the ability to use another type isn’t quite so obvious, and doing it “right” involves a tiny bit of extra work.

Or, consider that Lua could have descriptors, but the extra bit of work (especially design work) has been enough of an impediment that I’ve never implemented them. I don’t think the object implementations I’ve looked at have included them, either. Super weird!

In that light, it’s only natural that objects would be so strongly associated with the features Java and C++ attach to them. I think that makes it all the more important to play around! Look at what Moose has done. No, really, you should bear in mind my description of how Perl does stuff and flip through the Moose documentation. It’s amazing what they’ve built.

Amazon EC2 Bare Metal Instances with Direct Access to Hardware

Post Syndicated from Jeff Barr original https://aws.amazon.com/blogs/aws/new-amazon-ec2-bare-metal-instances-with-direct-access-to-hardware/

When customers come to us with new and unique requirements for AWS, we listen closely, ask lots of questions, and do our best to understand and address their needs. When we do this, we make the resulting service or feature generally available; we do not build one-offs or “snowflakes” for individual customers. That model is messy and hard to scale and is not the way we work.

Instead, every AWS customer has access to whatever it is that we build, and everyone benefits. VMware Cloud on AWS is a good example of this strategy in action. They told us that they wanted to run their virtualization stack directly on the hardware, within the AWS Cloud, giving their customers access to the elasticity, security, and reliability (not to mention the broad array of services) that AWS offers.

We knew that other customers also had interesting use cases for bare metal hardware and didn’t want to take the performance hit of nested virtualization. They wanted access to the physical resources for applications that take advantage of low-level hardware features such as performance counters and Intel® VT that are not always available or fully supported in virtualized environments, and also for applications intended to run directly on the hardware or licensed and supported for use in non-virtualized environments.

Our multi-year effort to move networking, storage, and other EC2 features out of our virtualization platform and into dedicated hardware was already well underway and provided the perfect foundation for a possible solution. This work, as I described in Now Available – Compute-Intensive C5 Instances for Amazon EC2, includes a set of dedicated hardware accelerators.

Now that we have provided VMware with the bare metal access that they requested, we are doing the same for all AWS customers. I’m really looking forward to seeing what you can do with them!

New Bare Metal Instances
Today we are launching a public preview the i3.metal instance, the first in a series of EC2 instances that offer the best of both worlds, allowing the operating system to run directly on the underlying hardware while still providing access to all of the benefits of the cloud. The instance gives you direct access to the processor and other hardware, and has the following specifications:

  • Processing – Two Intel Xeon E5-2686 v4 processors running at 2.3 GHz, with a total of 36 hyperthreaded cores (72 logical processors).
  • Memory – 512 GiB.
  • Storage – 15.2 terabytes of local, SSD-based NVMe storage.
  • Network – 25 Gbps of ENA-based enhanced networking.

Bare Metal instances are full-fledged members of the EC2 family and can take advantage of Elastic Load Balancing, Auto Scaling, Amazon CloudWatch, Auto Recovery, and so forth. They can also access the full suite of AWS database, IoT, mobile, analytics, artificial intelligence, and security services.

Previewing Now
We are launching a public preview of the Bare Metal instances today; please sign up now if you want to try them out.

You can now bring your specialized applications or your own stack of virtualized components to AWS and run them on Bare Metal instances. If you are using or thinking about using containers, these instances make a great host for CoreOS.

An AMI that works on one of the new C5 instances should also work on an I3 Bare Metal Instance. It must have the ENA and NVMe drivers, and must be tagged for ENA.

Jeff;

 

AWS PrivateLink Update – VPC Endpoints for Your Own Applications & Services

Post Syndicated from Jeff Barr original https://aws.amazon.com/blogs/aws/aws-privatelink-update-vpc-endpoints-for-your-own-applications-services/

Earlier this month, my colleague Colm MacCárthaigh told you about AWS PrivateLink and showed you how to use it to access AWS services such as Amazon Kinesis Streams, AWS Service Catalog, EC2 Systems Manager, the EC2 APIs, and the ELB APIs by way of VPC Endpoints. The endpoint (represented by one or more Elastic Network Interfaces or ENIs) resides within your VPC and has IP addresses drawn from the VPC’s subnets, without the need for an Internet or NAT Gateway. This model is clear and easy to understand, not to mention secure and scalable!

Endpoints for Private Connectivity
Today we are building upon the initial launch and extending the PrivateLink model, allowing you to set up and use VPC Endpoints to access your own services and those made available by others. Even before we launched PrivateLink for AWS services, we had a lot of requests for this feature, so I expect it to be pretty popular. For example, one customer told us that they plan to create hundreds of VPCs, each hosting and providing a single microservice (read Microservices on AWS to learn more).

Companies can now create services and offer them for sale to other AWS customers, for access via a private connection. They create a service that accepts TCP traffic, host it behind a Network Load Balancer, and then make the service available, either directly or in AWS Marketplace. They will be notified of new subscription requests and can choose to accept or reject each one. I expect that this feature will be used to create a strong, vibrant ecosystem of service providers in 2018.

The service provider and the service consumer run in separate VPCs and AWS accounts and communicate solely through the endpoint, with all traffic flowing across Amazon’s private network. Service consumers don’t have to worry about overlapping IP addresses, arrange for VPC peering, or use a VPC Gateway. You can also use AWS Direct Connect to connect your existing data center to one of your VPCs in order to allow your cloud-based applications to access services running on-premises, or vice versa.

Providing and Consuming Services
This new feature puts a lot of power at your fingertips. You can set it all up using the VPC APIs, the VPC CLI, or the AWS Management Console. I’ll use the console, and will show you how to provide and then consume a service. I am going to do both within a single AWS account, but that’s just for demo purposes.

Let’s talk about providing a service. It must run behind a Network Load Balancer and must be accessible over TCP. It can be hosted on EC2 instances, ECS containers, or on-premises (configured as an IP target), and should be able to scale in order to meet the expected level of demand. For low latency and fault tolerance, we recommend using an NLB with targets in every AZ of its region. Here’s mine:

I open up the VPC Console and navigate to Endpoint Services, then click on Create Endpoint Service:

I choose my NLB (just one in this case, but I can choose two or more and they will be mapped to consumers on a round-robin basis). By clicking on Acceptance required, I get to control access to my endpoint on a request-by-request basis:

I click on Create service and my service is ready immediately:

If I was going to make this service available in AWS Marketplace, I would go ahead and create a listing now. Since I am going to be the producer and the consumer in this blog post, I’ll skip that step. I will, however, copy the Service name for use in the next step.

I return to the VPC Dashboard and navigate to Endpoints, then click on Create endpoint. Then I select Find service by name, paste the service name, and click on Verify to move ahead. Then I select the desired AZs, and a subnet in each one, pick my security groups, and click on Create endpoint:

Because I checked Acceptance required when I created the endpoint service, the connection is pending acceptance:

Back on the endpoint service side (typically in a separate AWS account), I can see and accept the pending request:

The endpoint becomes available and ready to use within a minute or so. If I was creating a service and selling access on a paid basis, I would accept the request as part of a larger, and perhaps automated, onboarding workflow for a new customer.

On the consumer side, my new endpoint is accessible via DNS name:

Services provided by AWS and services in AWS Marketplace are accessible through split-horizon DNS. Accessing the service through this name will resolve to the “best” endpoint, taking Region and Availability Zone into consideration.

In the Marketplace
As I noted earlier, this new PrivateLink feature creates an opportunity for new and existing sellers in AWS Marketplace. The following SaaS offerings are already available as endpoints and I expect many more to follow (read Sell on AWS Marketplace to get started):

CA TechnologiesCA App Experience Analytics Essentials.

Aqua SecurityAqua Container Image Security Scanner.

DynatraceCloud-Native Monitoring powered by AI.

Cisco StealthwatchPublic Cloud Monitoring – Metered, Public Cloud Monitoring – Contracts.

SigOptML Optimization & Tuning.

Available Today
This new PrivateLink feature is available now and you can start using it today!

Jeff;

 

Raspberry Pi clusters come of age

Post Syndicated from Alex Bate original https://www.raspberrypi.org/blog/raspberry-pi-clusters-come-of-age/

In today’s guest post, Bruce Tulloch, CEO and Managing Director of BitScope Designs, discusses the uses of cluster computing with the Raspberry Pi, and the recent pilot of the Los Alamos National Laboratory 3000-Pi cluster built with the BitScope Blade.

Raspberry Pi cluster

High-performance computing and Raspberry Pi are not normally uttered in the same breath, but Los Alamos National Laboratory is building a Raspberry Pi cluster with 3000 cores as a pilot before scaling up to 40 000 cores or more next year.

That’s amazing, but why?

I was asked this question more than any other at The International Conference for High-Performance Computing, Networking, Storage and Analysis in Denver last week, where one of the Los Alamos Raspberry Pi Cluster Modules was on display at the University of New Mexico’s Center for Advanced Research Computing booth.

The short answer to this question is: the Raspberry Pi cluster enables Los Alamos National Laboratory (LANL) to conduct exascale computing R&D.

The Pi cluster breadboard

Exascale refers to computing systems at least 50 times faster than the most powerful supercomputers in use today. The problem faced by LANL and similar labs building these things is one of scale. To get the required performance, you need a lot of nodes, and to make it work, you need a lot of R&D.

However, there’s a catch-22: how do you write the operating systems, networks stacks, launch and boot systems for such large computers without having one on which to test it all? Use an existing supercomputer? No — the existing large clusters are fully booked 24/7 doing science, they cost millions of dollars per year to run, and they may not have the architecture you need for your next-generation machine anyway. Older machines retired from science may be available, but at this scale they cost far too much to use and are usually very hard to maintain.

The Los Alamos solution? Build a “model supercomputer” with Raspberry Pi!

Think of it as a “cluster development breadboard”.

The idea is to design, develop, debug, and test new network architectures and systems software on the “breadboard”, but at a scale equivalent to the production machines you’re currently building. Raspberry Pi may be a small computer, but it can run most of the system software stacks that production machines use, and the ratios of its CPU speed, local memory, and network bandwidth scale proportionately to the big machines, much like an architect’s model does when building a new house. To learn more about the project, see the news conference and this interview with insideHPC at SC17.

Traditional Raspberry Pi clusters

Like most people, we love a good cluster! People have been building them with Raspberry Pi since the beginning, because it’s inexpensive, educational, and fun. They’ve been built with the original Pi, Pi 2, Pi 3, and even the Pi Zero, but none of these clusters have proven to be particularly practical.

That’s not stopped them being useful though! I saw quite a few Raspberry Pi clusters at the conference last week.

One tiny one that caught my eye was from the people at openio.io, who used a small Raspberry Pi Zero W cluster to demonstrate their scalable software-defined object storage platform, which on big machines is used to manage petabytes of data, but which is so lightweight that it runs just fine on this:

Raspberry Pi Zero cluster

There was another appealing example at the ARM booth, where the Berkeley Labs’ singularity container platform was demonstrated running very effectively on a small cluster built with Raspberry Pi 3s.

Raspberry Pi 3 cluster demo at a conference stall

My show favourite was from the Edinburgh Parallel Computing Center (EPCC): Nick Brown used a cluster of Pi 3s to explain supercomputers to kids with an engaging interactive application. The idea was that visitors to the stand design an aircraft wing, simulate it across the cluster, and work out whether an aircraft that uses the new wing could fly from Edinburgh to New York on a full tank of fuel. Mine made it, fortunately!

Raspberry Pi 3 cluster demo at a conference stall

Next-generation Raspberry Pi clusters

We’ve been building small-scale industrial-strength Raspberry Pi clusters for a while now with BitScope Blade.

When Los Alamos National Laboratory approached us via HPC provider SICORP with a request to build a cluster comprising many thousands of nodes, we considered all the options very carefully. It needed to be dense, reliable, low-power, and easy to configure and to build. It did not need to “do science”, but it did need to work in almost every other way as a full-scale HPC cluster would.

Some people argue Compute Module 3 is the ideal cluster building block. It’s very small and just as powerful as Raspberry Pi 3, so one could, in theory, pack a lot of them into a very small space. However, there are very good reasons no one has ever successfully done this. For a start, you need to build your own network fabric and I/O, and cooling the CM3s, especially when densely packed in a cluster, is tricky given their tiny size. There’s very little room for heatsinks, and the tiny PCBs dissipate very little excess heat.

Instead, we saw the potential for Raspberry Pi 3 itself to be used to build “industrial-strength clusters” with BitScope Blade. It works best when the Pis are properly mounted, powered reliably, and cooled effectively. It’s important to avoid using micro SD cards and to connect the nodes using wired networks. It has the added benefit of coming with lots of “free” USB I/O, and the Pi 3 PCB, when mounted with the correct air-flow, is a remarkably good heatsink.

When Gordon announced netboot support, we became convinced the Raspberry Pi 3 was the ideal candidate when used with standard switches. We’d been making smaller clusters for a while, but netboot made larger ones practical. Assembling them all into compact units that fit into existing racks with multiple 10 Gb uplinks is the solution that meets LANL’s needs. This is a 60-node cluster pack with a pair of managed switches by Ubiquiti in testing in the BitScope Lab:

60-node Raspberry Pi cluster pack

Two of these packs, built with Blade Quattro, and one smaller one comprising 30 nodes, built with Blade Duo, are the components of the Cluster Module we exhibited at the show. Five of these modules are going into Los Alamos National Laboratory for their pilot as I write this.

Bruce Tulloch at a conference stand with a demo of the Raspberry Pi cluster for LANL

It’s not only research clusters like this for which Raspberry Pi is well suited. You can build very reliable local cloud computing and data centre solutions for research, education, and even some industrial applications. You’re not going to get much heavy-duty science, big data analytics, AI, or serious number crunching done on one of these, but it is quite amazing to see just how useful Raspberry Pi clusters can be for other purposes, whether it’s software-defined networks, lightweight MaaS, SaaS, PaaS, or FaaS solutions, distributed storage, edge computing, industrial IoT, and of course, education in all things cluster and parallel computing. For one live example, check out Mythic Beasts’ educational compute cloud, built with Raspberry Pi 3.

For more information about Raspberry Pi clusters, drop by BitScope Clusters.

I’ll read and respond to your thoughts in the comments below this post too.

Editor’s note:

Here is a photo of Bruce wearing a jetpack. Cool, right?!

Bruce Tulloch wearing a jetpack

The post Raspberry Pi clusters come of age appeared first on Raspberry Pi.

AWS Media Services – Process, Store, and Monetize Cloud-Based Video

Post Syndicated from Jeff Barr original https://aws.amazon.com/blogs/aws/aws-media-services-process-store-and-monetize-cloud-based-video/

Do you remember what web video was like in the early days? Standalone players, video no larger than a postage stamp, slow & cantankerous connections, overloaded servers, and the ever-present buffering messages were the norm less than two decades ago.

Today, thanks to technological progress and a broad array of standards, things are a lot better. Video consumers are now in control. They use devices of all shapes, sizes, and vintages to enjoy live and recorded content that is broadcast, streamed, or sent over-the-top (OTT, as they say), and expect immediate access to content that captures and then holds their attention. Meeting these expectations presents a challenge for content creators and distributors. Instead of generating video in a one-size-fits-all format, they (or their media servers) must be prepared to produce video that spans a broad range of sizes, formats, and bit rates, taking care to be ready to deal with planned or unplanned surges in demand. In the face of all of this complexity, they must backstop their content with a monetization model that supports the content and the infrastructure to deliver it.

New AWS Media Services
Today we are launching an array of broadcast-quality media services, each designed to address one or more aspects of the challenge that I outlined above. You can use them together to build a complete end-to-end video solution or you can use one or more in building-block style. In true AWS fashion, you can spend more time innovating and less time setting up and running infrastructure, leaving you ready to focus on creating, delivering, and monetizing your content. The services are all elastic, allowing you to ramp up processing power, connections, and storage and giving you the ability to handle million-user (and beyond) spikes with ease.

Here are the services (all accessible from a set of interactive consoles as well as through a comprehensive set of APIs):

AWS Elemental MediaConvert – File-based transcoding for OTT, broadcast, or archiving, with support for a long list of formats and codecs. Features include multi-channel audio, graphic overlays, closed captioning, and several DRM options.

AWS Elemental MediaLive – Live encoding to deliver video streams in real time to both televisions and multiscreen devices. Allows you to deploy highly reliable live channels in minutes, with full control over encoding parameters. It supports ad insertion, multi-channel audio, graphic overlays, and closed captioning.

AWS Elemental MediaPackage – Video origination and just-in-time packaging. Starting from a single input, produces output for multiple devices representing a long list of current and legacy formats. Supports multiple monetization models, time-shifted live streaming, ad insertion, DRM, and blackout management.

AWS Elemental MediaStore – Media-optimized storage that enables high performance and low latency applications such as live streaming, while taking advantage of the scale and durability of Amazon Simple Storage Service (S3).

AWS Elemental MediaTailor – Monetization service that supports ad serving and server-side ad insertion, a broad range of devices, transcoding, and accurate reporting of server-side and client-side ad insertion.

Instead of listing out all of the features in the sections below, I’ve simply included as many screen shots as possible with the expectation that this will give you a better sense of the rich set of features, parameters, and settings that you get with this set of services.

AWS Elemental MediaConvert
MediaConvert allows you to transcode content that is stored in files. You can process individual files or entire media libraries, or anything in-between. You simply create a conversion job that specifies the content and the desired outputs, and submit it to MediaConvert. There’s no software to install or patch and the service scales to meet your needs without affecting turnaround time or performance.

The MediaConvert Console lets you manage Output presets, Job templates, Queues, and Jobs:

You can use a built-in system preset or you can make one of your own. You have full control over the settings when you make your own:

Jobs templates are named, and produce one or more output groups. You can add a new group to a template with a click:

When everything is ready to go, you create a job and make some final selections, then click on Create:

Each account starts with a default queue for jobs, where incoming work is processed in parallel using all processing resources available to the account. Adding queues does not add processing resources, but does cause them to be apportioned across queues. You can temporarily pause one queue in order to devote more resources to the others. You can submit jobs to paused queues and you can also cancel any that have yet to start.

Pricing for this service is based on the amount of video that you process and the features that you use.

AWS Elemental MediaLive
This service is for live encoding, and can be run 24×7. MediaLive channels are deployed on redundant resources distributed in two physically separated Availability Zones in order to provide the reliability expected by our customers in the broadcast industry. You can specify your inputs and define your channels in the MediaLive Console:

After you create an Input, you create a Channel and attach it to the Input:

You have full control over the settings for each channel:

 

AWS Elemental MediaPackage
This service lets you deliver video to many devices from a single source. It focuses on protection and just-in-time packaging, giving you the ability to provide your users with the desired content on the device of their choice. You simply create a channel to get started:

Then you add one or more endpoints. Once again, plenty of options and full control, including a startover window and a time delay:

You find the input URL, user name, and password for your channel and route your live video stream to it for packaging:

AWS Elemental MediaStore
MediaStore offers the performance, consistency, and latency required for live and on-demand media delivery. Objects are written and read into a new “temporal” tier of object storage for a limited amount of time, then move silently into S3 for long-lived durability. You simply create a storage container to group your media content:

The container is available within a minute or so:

Like S3 buckets, MediaStore containers have access policies and no limits on the number of objects or storage capacity.

MediaStore helps you to take full advantage of S3 by managing the object key names so as to maximize storage and retrieval throughput, in accord with the Request Rate and Performance Considerations.

AWS Elemental MediaTailor
This service takes care of server-side ad insertion while providing a broadcast-quality viewer experience by transcoding ad assets on the fly. Your customer’s video player asks MediaTailor for a playlist. MediaTailor, in turn, calls your Ad Decision Server and returns a playlist that references the origin server for your original video and the ads recommended by the Ad Decision Server. The video player makes all of its requests to a single endpoint in order to ensure that client-side ad-blocking is ineffective. You simply create a MediaTailor Configuration:

Context information is passed to the Ad Decision Server in the URL:

Despite the length of this post I have barely scratched the surface of the AWS Media Services. Once AWS re:Invent is in the rear view mirror I hope to do a deep dive and show you how to use each of these services.

Available Now
The entire set of AWS Media Services is available now and you can start using them today! Pricing varies by service, but is built around a pay-as-you-go model.

Jeff;

UI Testing at Scale with AWS Lambda

Post Syndicated from Stas Neyman original https://aws.amazon.com/blogs/devops/ui-testing-at-scale-with-aws-lambda/

This is a guest blog post by Wes Couch and Kurt Waechter from the Blackboard Internal Product Development team about their experience using AWS Lambda.

One year ago, one of our UI test suites took hours to run. Last month, it took 16 minutes. Today, it takes 39 seconds. Here’s how we did it.

The backstory:

Blackboard is a global leader in delivering robust and innovative education software and services to clients in higher education, government, K12, and corporate training. We have a large product development team working across the globe in at least 10 different time zones, with an internal tools team providing support for quality and workflows. We have been using Selenium Webdriver to perform automated cross-browser UI testing since 2007. Because we are now practicing continuous delivery, the automated UI testing challenge has grown due to the faster release schedule. On top of that, every commit made to each branch triggers an execution of our automated UI test suite. If you have ever implemented an automated UI testing infrastructure, you know that it can be very challenging to scale and maintain. Although there are services that are useful for testing different browser/OS combinations, they don’t meet our scale needs.

It used to take three hours to synchronously run our functional UI suite, which revealed the obvious need for parallel execution. Previously, we used Mesos to orchestrate a Selenium Grid Docker container for each test run. This way, we were able to run eight concurrent threads for test execution, which took an average of 16 minutes. Although this setup is fine for a single workflow, the cracks started to show when we reached the scale required for Blackboard’s mature product lines. Going beyond eight concurrent sessions on a single container introduced performance problems that impact the reliability of tests (for example, issues in Webdriver or the browser popping up frequently). We tried Mesos and considered Kubernetes for Selenium Grid orchestration, but the answer to scaling a Selenium Grid was to think smaller, not larger. This led to our breakthrough with AWS Lambda.

The solution:

We started using AWS Lambda for UI testing because it doesn’t require costly infrastructure or countless man hours to maintain. The steps we outline in this blog post took one work day, from inception to implementation. By simply packaging the UI test suite into a Lambda function, we can execute these tests in parallel on a massive scale. We use a custom JUnit test runner that invokes the Lambda function with a request to run each test from the suite. The runner then aggregates the results returned from each Lambda test execution.

Selenium is the industry standard for testing UI at scale. Although there are other options to achieve the same thing in Lambda, we chose this mature suite of tools. Selenium is backed by Google, Firefox, and others to help the industry drive their browsers with code. This makes Lambda and Selenium a compelling stack for achieving UI testing at scale.

Making Chrome Run in Lambda

Currently, Chrome for Linux will not run in Lambda due to an absent mount point. By rebuilding Chrome with a slight modification, as Marco Lüthy originally demonstrated, you can run it inside Lambda anyway! It took about two hours to build the current master branch of Chromium to build on a c4.4xlarge. Unfortunately, the current version of ChromeDriver, 2.33, does not support any version of Chrome above 62, so we’ll be using Marco’s modified version of version 60 for the near future.

Required System Libraries

The Lambda runtime environment comes with a subset of common shared libraries. This means we need to include some extra libraries to get Chrome and ChromeDriver to work. Anything that exists in the java resources folder during compile time is included in the base directory of the compiled jar file. When this jar file is deployed to Lambda, it is placed in the /var/task/ directory. This allows us to simply place the libraries in the java resources folder under a folder named lib/ so they are right where they need to be when the Lambda function is invoked.

To get these libraries, create an EC2 instance and choose the Amazon Linux AMI.

Next, use ssh to connect to the server. After you connect to the new instance, search for the libraries to find their locations.

sudo find / -name libgconf-2.so.4
sudo find / -name libORBit-2.so.0

Now that you have the locations of the libraries, copy these files from the EC2 instance and place them in the java resources folder under lib/.

Packaging the Tests

To deploy the test suite to Lambda, we used a simple Gradle tool called ShadowJar, which is similar to the Maven Shade Plugin. It packages the libraries and dependencies inside the jar that is built. Usually test dependencies and sources aren’t included in a jar, but for this instance we want to include them. To include the test dependencies, add this section to the build.gradle file.

shadowJar {
   from sourceSets.test.output
   configurations = [project.configurations.testRuntime]
}

Deploying the Test Suite

Now that our tests are packaged with the dependencies in a jar, we need to get them into a running Lambda function. We use  simple SAM  templates to upload the packaged jar into S3, and then deploy it to Lambda with our settings.

{
   "AWSTemplateFormatVersion": "2010-09-09",
   "Transform": "AWS::Serverless-2016-10-31",
   "Resources": {
       "LambdaTestHandler": {
           "Type": "AWS::Serverless::Function",
           "Properties": {
               "CodeUri": "./build/libs/your-test-jar-all.jar",
               "Runtime": "java8",
               "Handler": "com.example.LambdaTestHandler::handleRequest",
               "Role": "<YourLambdaRoleArn>",
               "Timeout": 300,
               "MemorySize": 1536
           }
       }
   }
}

We use the maximum timeout available to ensure our tests have plenty of time to run. We also use the maximum memory size because this ensures our Lambda function can support Chrome and other resources required to run a UI test.

Specifying the handler is important because this class executes the desired test. The test handler should be able to receive a test class and method. With this information it will then execute the test and respond with the results.

public LambdaTestResult handleRequest(TestRequest testRequest, Context context) {
   LoggerContainer.LOGGER = new Logger(context.getLogger());
  
   BlockJUnit4ClassRunner runner = getRunnerForSingleTest(testRequest);
  
   Result result = new JUnitCore().run(runner);

   return new LambdaTestResult(result);
}

Creating a Lambda-Compatible ChromeDriver

We provide developers with an easily accessible ChromeDriver for local test writing and debugging. When we are running tests on AWS, we have configured ChromeDriver to run them in Lambda.

To configure ChromeDriver, we first need to tell ChromeDriver where to find the Chrome binary. Because we know that ChromeDriver is going to be unzipped into the root task directory, we should point the ChromeDriver configuration at that location.

The settings for getting ChromeDriver running are mostly related to Chrome, which must have its working directories pointed at the tmp/ folder.

Start with the default DesiredCapabilities for ChromeDriver, and then add the following settings to enable your ChromeDriver to start in Lambda.

public ChromeDriver createLambdaChromeDriver() {
   ChromeOptions options = new ChromeOptions();

   // Set the location of the chrome binary from the resources folder
   options.setBinary("/var/task/chrome");

   // Include these settings to allow Chrome to run in Lambda
   options.addArguments("--disable-gpu");
   options.addArguments("--headless");
   options.addArguments("--window-size=1366,768");
   options.addArguments("--single-process");
   options.addArguments("--no-sandbox");
   options.addArguments("--user-data-dir=/tmp/user-data");
   options.addArguments("--data-path=/tmp/data-path");
   options.addArguments("--homedir=/tmp");
   options.addArguments("--disk-cache-dir=/tmp/cache-dir");
  
   DesiredCapabilities desiredCapabilities = DesiredCapabilities.chrome();
   desiredCapabilities.setCapability(ChromeOptions.CAPABILITY, options);
  
   return new ChromeDriver(desiredCapabilities);
}

Executing Tests in Parallel

You can approach parallel test execution in Lambda in many different ways. Your approach depends on the structure and design of your test suite. For our solution, we implemented a custom test runner that uses reflection and JUnit libraries to create a list of test cases we want run. When we have the list, we create a TestRequest object to pass into the Lambda function that we have deployed. In this TestRequest, we place the class name, test method, and the test run identifier. When the Lambda function receives this TestRequest, our LambdaTestHandler generates and runs the JUnit test. After the test is complete, the test result is sent to the test runner. The test runner compiles a result after all of the tests are complete. By executing the same Lambda function multiple times with different test requests, we can effectively run the entire test suite in parallel.

To get screenshots and other test data, we pipe those files during test execution to an S3 bucket under the test run identifier prefix. When the tests are complete, we link the files to each test execution in the report generated from the test run. This lets us easily investigate test executions.

Pro Tip: Dynamically Loading Binaries

AWS Lambda has a limit of 250 MB of uncompressed space for packaged Lambda functions. Because we have libraries and other dependencies to our test suite, we hit this limit when we tried to upload a function that contained Chrome and ChromeDriver (~140 MB). This test suite was not originally intended to be used with Lambda. Otherwise, we would have scrutinized some of the included libraries. To get around this limit, we used the Lambda functions temporary directory, which allows up to 500 MB of space at runtime. Downloading these binaries at runtime moves some of that space requirement into the temporary directory. This allows more room for libraries and dependencies. You can do this by grabbing Chrome and ChromeDriver from an S3 bucket and marking them as executable using built-in Java libraries. If you take this route, be sure to point to the new location for these executables in order to create a ChromeDriver.

private static void downloadS3ObjectToExecutableFile(String key) throws IOException {
   File file = new File("/tmp/" + key);

   GetObjectRequest request = new GetObjectRequest("s3-bucket-name", key);

   FileUtils.copyInputStreamToFile(s3client.getObject(request).getObjectContent(), file);
   file.setExecutable(true);
}

Lambda-Selenium Project Source

We have compiled an open source example that you can grab from the Blackboard Github repository. Grab the code and try it out!

https://blackboard.github.io/lambda-selenium/

Conclusion

One year ago, one of our UI test suites took hours to run. Last month, it took 16 minutes. Today, it takes 39 seconds. Thanks to AWS Lambda, we can reduce our build times and perform automated UI testing at scale!

Access Resources in a VPC from AWS CodeBuild Builds

Post Syndicated from John Pignata original https://aws.amazon.com/blogs/devops/access-resources-in-a-vpc-from-aws-codebuild-builds/

John Pignata, Startup Solutions Architect, Amazon Web Services

In this blog post we’re going to discuss a new AWS CodeBuild feature that is available starting today. CodeBuild builds can now access resources in a VPC directly without these resources being exposed to the public internet. These resources include Amazon Relational Database Service (Amazon RDS) databases, Amazon ElastiCache clusters, internal services running on Amazon Elastic Compute Cloud (Amazon EC2), and Amazon EC2 Container Service (Amazon ECS), or any service endpoints that are only reachable from within a specific VPC.

CodeBuild is a fully managed build service that compiles source code, runs tests, and produces software packages that are ready to deploy. As part of the build process, developers often require access to resources that should be isolated from the public Internet. Now CodeBuild builds can be optionally configured to have VPC connectivity and access these resources directly.

Accessing Resources in a VPC

You can configure builds to have access to a VPC when you create a CodeBuild project or you can update an existing CodeBuild project with VPC configuration attributes. Here’s how it looks in the console:

 

To configure VPC connectivity: select a VPC, one or more subnets within that VPC, and one or more VPC security groups that CodeBuild should apply when attaching to your VPC. Once configured, commands running as part of your build will be able to access resources in your VPC without transiting across the public Internet.

Use Cases

The availability of VPC connectivity from CodeBuild builds unlocks many potential uses. For example, you can:

  • Run integration tests from your build against data in an Amazon RDS instance that’s isolated on a private subnet.
  • Query data in an ElastiCache cluster directly from tests.
  • Interact with internal web services hosted on Amazon EC2, Amazon ECS, or services that use internal Elastic Load Balancing.
  • Retrieve dependencies from self-hosted, internal artifact repositories such as PyPI for Python, Maven for Java, npm for Node.js, and so on.
  • Access objects in an Amazon S3 bucket configured to allow access only through a VPC endpoint.
  • Query external web services that require fixed IP addresses through the Elastic IP address of the NAT gateway associated with your subnet(s).

… and more! Your builds can now access any resource that’s hosted in your VPC without any compromise on network isolation.

Internet Connectivity

CodeBuild requires access to resources on the public Internet to successfully execute builds. At a minimum, it must be able to reach your source repository system (such as AWS CodeCommit, GitHub, Bitbucket), Amazon Simple Storage Service (Amazon S3) to deliver build artifacts, and Amazon CloudWatch Logs to stream logs from the build process. The interface attached to your VPC will not be assigned a public IP address so to enable Internet access from your builds, you will need to set up a managed NAT Gateway or NAT instance for the subnets you configure. You must also ensure your security groups allow outbound access to these services.

IP Address Space

Each running build will be assigned an IP address from one of the subnets in your VPC that you designate for CodeBuild to use. As CodeBuild scales to meet your build volume, ensure that you select subnets with enough address space to accommodate your expected number of concurrent builds.

Service Role Permissions

CodeBuild requires new permissions in order to manage network interfaces on your VPCs. If you create a service role for your new projects, these permissions will be included in that role’s policy automatically. For existing service roles, you can edit the policy document to include the additional actions. For the full policy document to apply to your service role, see Advanced Setup in the CodeBuild documentation.

For more information, see VPC Support in the CodeBuild documentation. We hope you find the ability to access internal resources on a VPC useful in your build processes! If you have any questions or feedback, feel free to reach out to us through the AWS CodeBuild forum or leave a comment!

Introducing container-diff, a tool for quickly comparing container images (Google Open Source Blog)

Post Syndicated from jake original https://lwn.net/Articles/739384/rss

Google has announced that it has released its container-diff tool under the Apache v2 license.
container-diff helps users investigate image changes by computing semantic diffs between images. What this means is that container-diff figures out on a low-level what data changed, and then combines this with an understanding of package manager information to output this information in a format that’s actually readable to users. The tool can find differences in system packages, language-level packages, and files in a container image.

Users can specify images in several formats – from local Docker daemon (using the prefix `daemon://` on the image path), a remote registry (using the prefix `remote://`), or a file in the .tar in the format exported by “docker save” command. You can also combine these formats to compute the diff between a local version of an image and a remote version.”

timeShift(GrafanaBuzz, 1w) Issue 22

Post Syndicated from Blogs on Grafana Labs Blog original https://grafana.com/blog/2017/11/17/timeshiftgrafanabuzz-1w-issue-22/

Welome to TimeShift

We hope you liked our recent article with videos and slides from the events we’ve participated in recently. With Thanksgiving right around the corner, we’re getting a breather from work-related travel, but only a short one. We have some events in the coming weeks, and of course are busy filling in the details for GrafanaCon EU.

This week we have a lot of articles, videos and presentations to share, as well as some important plugin updates. Enjoy!


Latest Release

Grafana 4.6.2 is now available and includes some bug fixes:

  • Prometheus: Fixes bug with new Prometheus alerts in Grafana. Make sure to download this version if your using Prometheus for alerting. More details in the issue. #9777
  • Color picker: Bug after using textbox input field to change/paste color string #9769
  • Cloudwatch: build using golang 1.9.2 #9667, thanks @mtanda
  • Heatmap: Fixed tooltip for “time series buckets” mode #9332
  • InfluxDB: Fixed query editor issue when using > or < operators in WHERE clause #9871

Download Grafana 4.6.2 Now


From the Blogosphere

Cloud Tech 10 – 13th November 2017 – Grafana, Linux FUSE Adapter, Azure Stack and more!: Mark Whitby is a Cloud Solution Architect at Microsoft UK. Each week he prodcues a video reviewing new developments with Microsoft Azure. This week Mark covers the new Azure Monitoring Plugin we recently announced. He also shows you how to get up and running with Grafana quickly using the Azure Marketplace.

Using Prometheus and Grafana to Monitor WebLogic Server on Kubernetes: Oracle published an article on monitoring WebLogic server on Kubernetes. To do this, you’ll use the WebLogic Monitoring Exporter to scrape the server metrics and feed them to Prometheus, then visualize the data in Grafana. Marina goes into a lot of detail and provides sample files and configs to help you get going.

Getting Started with Prometheus: Will Robinson has started a new series on monitoring with Prometheus from someone who has never touched it before. Part 1 introduces a number of monitoring tools and concepts, and helps define a number of monitoring terms. Part 2 teaches you how to spin up Prometheus in a Docker container, and takes a look at writing queries. Looking forward to the third post, when he dives into the visualization aspect.

Monitoring with Prometheus: Alexander Schwartz has made the slides from his most recent presentation from the Continuous Lifcycle Conference in Germany available. In his talk, he discussed getting started with Prometheus, how it differs from other monitoring concepts, and provides examples of how to monitor and alert. We’ll link to the video of the talk when it’s available.

Using Grafana with SiriDB: Jeroen van der Heijden has written an in-depth tutorial to help you visualize data from the open source TSDB, SiriDB in Grafana. This tutorial will get you familiar with setting up SiriDB and provides a sample dashboard to help you get started.

Real-Time Monitoring with Grafana, StatsD and InfluxDB – Artur Caliendo Prado: This is a video from a talk at The Conf, held in Brazil. Artur’s presentation focuses on the experiences they had building a monitoring stack at Youse, how their monitoring became more complex as they scaled, and the platform they built to make sense of their data.

Using Grafana & Inlfuxdb to view XIV Host Performance Metrics – Part 4 Array Stats: This is the fourth part in a series of posts about host performance metrics. This post dives in to array stats to identify workloads and maintain balance across ports. Check out part 1, part 2 and part 3.


GrafanaCon Tickets are Going Fast

Tickets are going fast for GrafanaCon EU, but we still have a seat reserved for you. Join us March 1-2, 2018 in Amsterdam for 2 days of talks centered around Grafana and the surrounding monitoring ecosystem including Graphite, Prometheus, InfluxData, Elasticsearch, Kubernetes, and more.

Get Your Ticket Now


Grafana Plugins

Plugin authors are often adding new features and fixing bugs, which will make your plugin perform better – so it’s important to keep your plugins up to date. We’ve made updating easy; for on-prem Grafana, use the Grafana-cli tool, or update with 1 click if you’re using Hosted Grafana.

UPDATED PLUGIN

Hawkular data source – There is an important change in this release – as this datasource is now able to fetch not only Hawkular Metrics but also Hawkular Alerts, the server URL in the datasource configuration must be updated: http://myserver:123/hawkular/metrics must be changed to http://myserver:123/hawkular

Some of the changes (see the release notes) for more details):

  • Allow per-query tenant configuration
  • Annotations can now be configured out of Availability metrics and Hawkular Alerts events in addition to string metrics
  • allows dot character in tag names

Update

UPDATED PLUGIN

Diagram Panel – This is the first release in a while for the popular Diagram Panel plugin.

In addition to these changes, there are also a number of bug fixes:

Update

UPDATED PLUGIN

Influx Admin Panel – received a number of improvements:

  • Fix issue always showing query results
  • When there is only one row, swap rows/cols (ie: SHOW DIAGNOSTICS)
  • Improved auto-refresh behavior
  • Fix query time sorting
  • show ‘status’ field (killed, etc)

Update


Upcoming Events:

In between code pushes we like to speak at, sponsor and attend all kinds of conferences and meetups. We have some awesome talks and events coming soon. Hope to see you at one of these!

How to Use Open Source Projects for Performance Monitoring | Webinar
Nov. 29, 1pm EST
:
Check out how you can use popular open source projects, for performance monitoring of your Infrastructure, Application, and Cloud faster, easier, and to scale. In this webinar, Daniel Lee from Grafana Labs, and Chris Churilo from InfluxData, will provide you with step by step instruction from download & configure, to collecting metrics and building dashboards and alerts.

RSVP

KubeCon | Austin, TX – Dec. 6-8, 2017: We’re sponsoring KubeCon 2017! This is the must-attend conference for cloud native computing professionals. KubeCon + CloudNativeCon brings together leading contributors in:

  • Cloud native applications and computing
  • Containers
  • Microservices
  • Central orchestration processing
  • And more

Buy Tickets

FOSDEM | Brussels, Belgium – Feb 3-4, 2018: FOSDEM is a free developer conference where thousands of developers of free and open source software gather to share ideas and technology. Carl Bergquist is managing the Cloud and Monitoring Devroom, and the CFP is now open. There is no need to register; all are welcome. If you’re interested in speaking at FOSDEM, submit your talk now!


Tweet of the Week

We scour Twitter each week to find an interesting/beautiful dashboard and show it off! #monitoringLove

We were glad to be a part of InfluxDays this year, and looking forward to seeing the InfluxData team in NYC in February.


Grafana Labs is Hiring!

We are passionate about open source software and thrive on tackling complex challenges to build the future. We ship code from every corner of the globe and love working with the community. If this sounds exciting, you’re in luck – WE’RE HIRING!

Check out our Open Positions


How are we doing?

I enjoy writing these weekly roudups, but am curious how I can improve them. Submit a comment on this article below, or post something at our community forum. Help us make these weekly roundups better!

Follow us on Twitter, like us on Facebook, and join the Grafana Labs community.

Event-Driven Computing with Amazon SNS and AWS Compute, Storage, Database, and Networking Services

Post Syndicated from Christie Gifrin original https://aws.amazon.com/blogs/compute/event-driven-computing-with-amazon-sns-compute-storage-database-and-networking-services/

Contributed by Otavio Ferreira, Manager, Software Development, AWS Messaging

Like other developers around the world, you may be tackling increasingly complex business problems. A key success factor, in that case, is the ability to break down a large project scope into smaller, more manageable components. A service-oriented architecture guides you toward designing systems as a collection of loosely coupled, independently scaled, and highly reusable services. Microservices take this even further. To improve performance and scalability, they promote fine-grained interfaces and lightweight protocols.

However, the communication among isolated microservices can be challenging. Services are often deployed onto independent servers and don’t share any compute or storage resources. Also, you should avoid hard dependencies among microservices, to preserve maintainability and reusability.

If you apply the pub/sub design pattern, you can effortlessly decouple and independently scale out your microservices and serverless architectures. A pub/sub messaging service, such as Amazon SNS, promotes event-driven computing that statically decouples event publishers from subscribers, while dynamically allowing for the exchange of messages between them. An event-driven architecture also introduces the responsiveness needed to deal with complex problems, which are often unpredictable and asynchronous.

What is event-driven computing?

Given the context of microservices, event-driven computing is a model in which subscriber services automatically perform work in response to events triggered by publisher services. This paradigm can be applied to automate workflows while decoupling the services that collectively and independently work to fulfil these workflows. Amazon SNS is an event-driven computing hub, in the AWS Cloud, that has native integration with several AWS publisher and subscriber services.

Which AWS services publish events to SNS natively?

Several AWS services have been integrated as SNS publishers and, therefore, can natively trigger event-driven computing for a variety of use cases. In this post, I specifically cover AWS compute, storage, database, and networking services, as depicted below.

Compute services

  • Auto Scaling: Helps you ensure that you have the correct number of Amazon EC2 instances available to handle the load for your application. You can configure Auto Scaling lifecycle hooks to trigger events, as Auto Scaling resizes your EC2 cluster.As an example, you may want to warm up the local cache store on newly launched EC2 instances, and also download log files from other EC2 instances that are about to be terminated. To make this happen, set an SNS topic as your Auto Scaling group’s notification target, then subscribe two Lambda functions to this SNS topic. The first function is responsible for handling scale-out events (to warm up cache upon provisioning), whereas the second is in charge of handling scale-in events (to download logs upon termination).

  • AWS Elastic Beanstalk: An easy-to-use service for deploying and scaling web applications and web services developed in a number of programming languages. You can configure event notifications for your Elastic Beanstalk environment so that notable events can be automatically published to an SNS topic, then pushed to topic subscribers.As an example, you may use this event-driven architecture to coordinate your continuous integration pipeline (such as Jenkins CI). That way, whenever an environment is created, Elastic Beanstalk publishes this event to an SNS topic, which triggers a subscribing Lambda function, which then kicks off a CI job against your newly created Elastic Beanstalk environment.

  • Elastic Load Balancing: Automatically distributes incoming application traffic across Amazon EC2 instances, containers, or other resources identified by IP addresses.You can configure CloudWatch alarms on Elastic Load Balancing metrics, to automate the handling of events derived from Classic Load Balancers. As an example, you may leverage this event-driven design to automate latency profiling in an Amazon ECS cluster behind a Classic Load Balancer. In this example, whenever your ECS cluster breaches your load balancer latency threshold, an event is posted by CloudWatch to an SNS topic, which then triggers a subscribing Lambda function. This function runs a task on your ECS cluster to trigger a latency profiling tool, hosted on the cluster itself. This can enhance your latency troubleshooting exercise by making it timely.

Storage services

  • Amazon S3: Object storage built to store and retrieve any amount of data.You can enable S3 event notifications, and automatically get them posted to SNS topics, to automate a variety of workflows. For instance, imagine that you have an S3 bucket to store incoming resumes from candidates, and a fleet of EC2 instances to encode these resumes from their original format (such as Word or text) into a portable format (such as PDF).In this example, whenever new files are uploaded to your input bucket, S3 publishes these events to an SNS topic, which in turn pushes these messages into subscribing SQS queues. Then, encoding workers running on EC2 instances poll these messages from the SQS queues; retrieve the original files from the input S3 bucket; encode them into PDF; and finally store them in an output S3 bucket.

  • Amazon EFS: Provides simple and scalable file storage, for use with Amazon EC2 instances, in the AWS Cloud.You can configure CloudWatch alarms on EFS metrics, to automate the management of your EFS systems. For example, consider a highly parallelized genomics analysis application that runs against an EFS system. By default, this file system is instantiated on the “General Purpose” performance mode. Although this performance mode allows for lower latency, it might eventually impose a scaling bottleneck. Therefore, you may leverage an event-driven design to handle it automatically.Basically, as soon as the EFS metric “Percent I/O Limit” breaches 95%, CloudWatch could post this event to an SNS topic, which in turn would push this message into a subscribing Lambda function. This function automatically creates a new file system, this time on the “Max I/O” performance mode, then switches the genomics analysis application to this new file system. As a result, your application starts experiencing higher I/O throughput rates.

  • Amazon Glacier: A secure, durable, and low-cost cloud storage service for data archiving and long-term backup.You can set a notification configuration on an Amazon Glacier vault so that when a job completes, a message is published to an SNS topic. Retrieving an archive from Amazon Glacier is a two-step asynchronous operation, in which you first initiate a job, and then download the output after the job completes. Therefore, SNS helps you eliminate polling your Amazon Glacier vault to check whether your job has been completed, or not. As usual, you may subscribe SQS queues, Lambda functions, and HTTP endpoints to your SNS topic, to be notified when your Amazon Glacier job is done.

  • AWS Snowball: A petabyte-scale data transport solution that uses secure appliances to transfer large amounts of data.You can leverage Snowball notifications to automate workflows related to importing data into and exporting data from AWS. More specifically, whenever your Snowball job status changes, Snowball can publish this event to an SNS topic, which in turn can broadcast the event to all its subscribers.As an example, imagine a Geographic Information System (GIS) that distributes high-resolution satellite images to users via Web browser. In this example, the GIS vendor could capture up to 80 TB of satellite images; create a Snowball job to import these files from an on-premises system to an S3 bucket; and provide an SNS topic ARN to be notified upon job status changes in Snowball. After Snowball changes the job status from “Importing” to “Completed”, Snowball publishes this event to the specified SNS topic, which delivers this message to a subscribing Lambda function, which finally creates a CloudFront web distribution for the target S3 bucket, to serve the images to end users.

Database services

  • Amazon RDS: Makes it easy to set up, operate, and scale a relational database in the cloud.RDS leverages SNS to broadcast notifications when RDS events occur. As usual, these notifications can be delivered via any protocol supported by SNS, including SQS queues, Lambda functions, and HTTP endpoints.As an example, imagine that you own a social network website that has experienced organic growth, and needs to scale its compute and database resources on demand. In this case, you could provide an SNS topic to listen to RDS DB instance events. When the “Low Storage” event is published to the topic, SNS pushes this event to a subscribing Lambda function, which in turn leverages the RDS API to increase the storage capacity allocated to your DB instance. The provisioning itself takes place within the specified DB maintenance window.

  • Amazon ElastiCache: A web service that makes it easy to deploy, operate, and scale an in-memory data store or cache in the cloud.ElastiCache can publish messages using Amazon SNS when significant events happen on your cache cluster. This feature can be used to refresh the list of servers on client machines connected to individual cache node endpoints of a cache cluster. For instance, an ecommerce website fetches product details from a cache cluster, with the goal of offloading a relational database and speeding up page load times. Ideally, you want to make sure that each web server always has an updated list of cache servers to which to connect.To automate this node discovery process, you can get your ElastiCache cluster to publish events to an SNS topic. Thus, when ElastiCache event “AddCacheNodeComplete” is published, your topic then pushes this event to all subscribing HTTP endpoints that serve your ecommerce website, so that these HTTP servers can update their list of cache nodes.

  • Amazon Redshift: A fully managed data warehouse that makes it simple to analyze data using standard SQL and BI (Business Intelligence) tools.Amazon Redshift uses SNS to broadcast relevant events so that data warehouse workflows can be automated. As an example, imagine a news website that sends clickstream data to a Kinesis Firehose stream, which then loads the data into Amazon Redshift, so that popular news and reading preferences might be surfaced on a BI tool. At some point though, this Amazon Redshift cluster might need to be resized, and the cluster enters a ready-only mode. Hence, this Amazon Redshift event is published to an SNS topic, which delivers this event to a subscribing Lambda function, which finally deletes the corresponding Kinesis Firehose delivery stream, so that clickstream data uploads can be put on hold.At a later point, after Amazon Redshift publishes the event that the maintenance window has been closed, SNS notifies a subscribing Lambda function accordingly, so that this function can re-create the Kinesis Firehose delivery stream, and resume clickstream data uploads to Amazon Redshift.

  • AWS DMS: Helps you migrate databases to AWS quickly and securely. The source database remains fully operational during the migration, minimizing downtime to applications that rely on the database.DMS also uses SNS to provide notifications when DMS events occur, which can automate database migration workflows. As an example, you might create data replication tasks to migrate an on-premises MS SQL database, composed of multiple tables, to MySQL. Thus, if replication tasks fail due to incompatible data encoding in the source tables, these events can be published to an SNS topic, which can push these messages into a subscribing SQS queue. Then, encoders running on EC2 can poll these messages from the SQS queue, encode the source tables into a compatible character set, and restart the corresponding replication tasks in DMS. This is an event-driven approach to a self-healing database migration process.

Networking services

  • Amazon Route 53: A highly available and scalable cloud-based DNS (Domain Name System). Route 53 health checks monitor the health and performance of your web applications, web servers, and other resources.You can set CloudWatch alarms and get automated Amazon SNS notifications when the status of your Route 53 health check changes. As an example, imagine an online payment gateway that reports the health of its platform to merchants worldwide, via a status page. This page is hosted on EC2 and fetches platform health data from DynamoDB. In this case, you could configure a CloudWatch alarm for your Route 53 health check, so that when the alarm threshold is breached, and the payment gateway is no longer considered healthy, then CloudWatch publishes this event to an SNS topic, which pushes this message to a subscribing Lambda function, which finally updates the DynamoDB table that populates the status page. This event-driven approach avoids any kind of manual update to the status page visited by merchants.

  • AWS Direct Connect (AWS DX): Makes it easy to establish a dedicated network connection from your premises to AWS, which can reduce your network costs, increase bandwidth throughput, and provide a more consistent network experience than Internet-based connections.You can monitor physical DX connections using CloudWatch alarms, and send SNS messages when alarms change their status. As an example, when a DX connection state shifts to 0 (zero), indicating that the connection is down, this event can be published to an SNS topic, which can fan out this message to impacted servers through HTTP endpoints, so that they might reroute their traffic through a different connection instead. This is an event-driven approach to connectivity resilience.

More event-driven computing on AWS

In addition to SNS, event-driven computing is also addressed by Amazon CloudWatch Events, which delivers a near real-time stream of system events that describe changes in AWS resources. With CloudWatch Events, you can route each event type to one or more targets, including:

Many AWS services publish events to CloudWatch. As an example, you can get CloudWatch Events to capture events on your ETL (Extract, Transform, Load) jobs running on AWS Glue and push failed ones to an SQS queue, so that you can retry them later.

Conclusion

Amazon SNS is a pub/sub messaging service that can be used as an event-driven computing hub to AWS customers worldwide. By capturing events natively triggered by AWS services, such as EC2, S3 and RDS, you can automate and optimize all kinds of workflows, namely scaling, testing, encoding, profiling, broadcasting, discovery, failover, and much more. Business use cases presented in this post ranged from recruiting websites, to scientific research, geographic systems, social networks, retail websites, and news portals.

Start now by visiting Amazon SNS in the AWS Management Console, or by trying the AWS 10-Minute Tutorial, Send Fan-out Event Notifications with Amazon SNS and Amazon SQS.