Tag Archives: Dyn

Introducing Cost Allocation Tags for Amazon SQS

Post Syndicated from Jeff Barr original https://aws.amazon.com/blogs/aws/introducing-cost-allocation-tags-for-amazon-sqs/

You have long had the ability to tag your AWS resources and to see cost breakouts on a per-tag basis. Cost allocation was launched in 2012 (see AWS Cost Allocation for Customer Bills) and we have steadily added support for additional services, most recently DynamoDB (Introducing Cost Allocation Tags for Amazon DynamoDB), Lambda (AWS Lambda Supports Tagging and Cost Allocations), and EBS (New – Cost Allocation for AWS Snapshots).

Today, we are launching tag-based cost allocation for Amazon Simple Queue Service (SQS). You can now assign tags to your queues and use them to manage your costs at any desired level: application, application stage (for a loosely coupled application that communicates via queues), project, department, or developer. After you have tagged your queues, you can use the AWS Tag Editor to search queues that have tags of interest.

Here’s how I would add three tags (app, stage, and department) to one of my queues:

This feature is available now in all AWS Regions and you can start using in today! To learn more about tagging, read Tagging Your Amazon SQS Queues. To learn more about cost allocation via tags, read Using Cost Allocation Tags. To learn more about how to use message queues to build loosely coupled microservices for modern applications, read our blog post (Building Loosely Coupled, Scalable, C# Applications with Amazon SQS and Amazon SNS) and watch the recording of our recent webinar, Decouple and Scale Applications Using Amazon SQS and Amazon SNS.

If you are coming to AWS re:Invent, plan to attend session ARC 330: How the BBC Built a Massive Media Pipeline Using Microservices. In the talk you will find out how they used SNS and SQS to improve the elasticity and reliability of the BBC iPlayer architecture.

Jeff;

Coaxing 2D platforming out of Unity

Post Syndicated from Eevee original https://eev.ee/blog/2017/10/13/coaxing-2d-platforming-out-of-unity/

An anonymous donor asked a question that I can’t even begin to figure out how to answer, but they also said anything else is fine, so here’s anything else.

I’ve been avoiding writing about game physics, since I want to save it for ✨ the book I’m writing ✨, but that book will almost certainly not touch on Unity. Here, then, is a brief run through some of the brick walls I ran into while trying to convince Unity to do 2D platforming.

This is fairly high-level — there are no blocks of code or helpful diagrams. I’m just getting this out of my head because it’s interesting. If you want more gritty details, I guess you’ll have to wait for ✨ the book ✨.

The setup

I hadn’t used Unity before. I hadn’t even used a “real” physics engine before. My games so far have mostly used LÖVE, a Lua-based engine. LÖVE includes box2d bindings, but for various reasons (not all of them good), I opted to avoid them and instead write my own physics completely from scratch. (How, you ask? ✨ Book ✨!)

I was invited to work on a Unity project, Chaos Composer, that someone else had already started. It had basic movement already implemented; I taught myself Unity’s physics system by hacking on it. It’s entirely possible that none of this is actually the best way to do anything, since I was really trying to reproduce my own homegrown stuff in Unity, but it’s the best I’ve managed to come up with.

Two recurring snags were that you can’t ask Unity to do multiple physics updates in a row, and sometimes getting the information I wanted was difficult. Working with my own code spoiled me a little, since I could invoke it at any time and ask it anything I wanted; Unity, on the other hand, is someone else’s black box with a rigid interface on top.

Also, wow, Googling for a lot of this was not quite as helpful as expected. A lot of what’s out there is just the first thing that works, and often that’s pretty hacky and imposes severe limits on the game design (e.g., “this won’t work with slopes”). Basic movement and collision are the first thing you do, which seems to me like the worst time to be locking yourself out of a lot of design options. I tried very (very, very, very) hard to minimize those kinds of constraints.

Problem 1: Movement

When I showed up, movement was already working. Problem solved!

Like any good programmer, I immediately set out to un-solve it. Given a “real” physics engine like Unity prominently features, you have two options: ⓐ treat the player as a physics object, or ⓑ don’t. The existing code went with option ⓑ, like I’d done myself with LÖVE, and like I’d seen countless people advise. Using a physics sim makes for bad platforming.

But… why? I believed it, but I couldn’t concretely defend it. I had to know for myself. So I started a blank project, drew some physics boxes, and wrote a dozen-line player controller.

Ah! Immediate enlightenment.

If the player was sliding down a wall, and I tried to move them into the wall, they would simply freeze in midair until I let go of the movement key. The trouble is that the physics sim works in terms of forces — moving the player involves giving them a nudge in some direction, like a giant invisible hand pushing them around the level. Surprise! If you press a real object against a real wall with your real hand, you’ll see the same effect — friction will cancel out gravity, and the object will stay in midair..

Platformer movement, as it turns out, doesn’t make any goddamn physical sense. What is air control? What are you pushing against? Nothing, really; we just have it because it’s nice to play with, because not having it is a nightmare.

I looked to see if there were any common solutions to this, and I only really found one: make all your walls frictionless.

Game development is full of hacks like this, and I… don’t like them. I can accept that minor hacks are necessary sometimes, but this one makes an early and widespread change to a fundamental system to “fix” something that was wrong in the first place. It also imposes an “invisible” requirement, something I try to avoid at all costs — if you forget to make a particular wall frictionless, you’ll never know unless you happen to try sliding down it.

And so, I swiftly returned to the existing code. It wasn’t too different from what I’d come up with for LÖVE: it applied gravity by hand, tracked the player’s velocity, computed the intended movement each frame, and moved by that amount. The interesting thing was that it used MovePosition, which schedules a movement for the next physics update and stops the movement if the player hits something solid.

It’s kind of a nice hybrid approach, actually; all the “physics” for conscious actors is done by hand, but the physics engine is still used for collision detection. It’s also used for collision rejection — if the player manages to wedge themselves several pixels into a solid object, for example, the physics engine will try to gently nudge them back out of it with no extra effort required on my part. I still haven’t figured out how to get that to work with my homegrown stuff, which is built to prevent overlap rather than to jiggle things out of it.

But wait, what about…

Our player is a dynamic body with rotation lock and no gravity. Why not just use a kinematic body?

I must be missing something, because I do not understand the point of kinematic bodies. I ran into this with Godot, too, which documented them the same way: as intended for use as players and other manually-moved objects. But by default, they don’t even collide with other kinematic bodies or static geometry. What? There’s a checkbox to turn this on, which I enabled, but then I found out that MovePosition doesn’t stop kinematic bodies when they hit something, so I would’ve had to cast along the intended path of movement to figure out when to stop, thus duplicating the same work the physics engine was about to do.

But that’s impossible anyway! Static geometry generally wants to be made of edge colliders, right? They don’t care about concave/convex. Imagine the player is standing on the ground near a wall and tries to move towards the wall. Both the ground and the wall are different edges from the same edge collider.

If you try to cast the player’s hitbox horizontally, parallel to the ground, you’ll only get one collision: the existing collision with the ground. Casting doesn’t distinguish between touching and hitting. And because Unity only reports one collision per collider, and because the ground will always show up first, you will never find out about the impending wall collision.

So you’re forced to either use raycasts for collision detection or decomposed polygons for world geometry, both of which are slightly worse tools for no real gain.

I ended up sticking with a dynamic body.


Oh, one other thing that doesn’t really fit anywhere else: keep track of units! If you’re adding something called “velocity” directly to something called “position”, something has gone very wrong. Acceleration is distance per time squared; velocity is distance per time; position is distance. You must multiply or divide by time to convert between them.

I never even, say, add a constant directly to position every frame; I always phrase it as velocity and multiply by Δt. It keeps the units consistent: time is always in seconds, not in tics.

Problem 2: Slopes

Ah, now we start to get off in the weeds.

A sort of pre-problem here was detecting whether we’re on a slope, which means detecting the ground. The codebase originally used a manual physics query of the area around the player’s feet to check for the ground, which seems to be somewhat common, but that can’t tell me the angle of the detected ground. (It’s also kind of error-prone, since “around the player’s feet” has to be specified by hand and may not stay correct through animations or changes in the hitbox.)

I replaced that with what I’d eventually settled on in LÖVE: detect the ground by detecting collisions, and looking at the normal of the collision. A normal is a vector that points straight out from a surface, so if you’re standing on the ground, the normal points straight up; if you’re on a 10° incline, the normal points 10° away from straight up.

Not all collisions are with the ground, of course, so I assumed something is ground if the normal pointed away from gravity. (I like this definition more than “points upwards”, because it avoids assuming anything about the direction of gravity, which leaves some interesting doors open for later on.) That’s easily detected by taking the dot product — if it’s negative, the collision was with the ground, and I now have the normal of the ground.

Actually doing this in practice was slightly tricky. With my LÖVE engine, I could cram this right into the middle of collision resolution. With Unity, not quite so much. I went through a couple iterations before I really grasped Unity’s execution order, which I guess I will have to briefly recap for this to make sense.

Unity essentially has two update cycles. It performs physics updates at fixed intervals for consistency, and updates everything else just before rendering. Within a single frame, Unity does as many fixed physics updates as it has spare time for (which might be zero, one, or more), then does a regular update, then renders. User code can implement either or both of Update, which runs during a regular update, and FixedUpdate, which runs just before Unity does a physics pass.

So my solution was:

  • At the very end of FixedUpdate, clear the actor’s “on ground” flag and ground normal.

  • During OnCollisionEnter2D and OnCollisionStay2D (which are called from within a physics pass), if there’s a collision that looks like it’s with the ground, set the “on ground” flag and ground normal. (If there are multiple ground collisions, well, good luck figuring out the best way to resolve that! At the moment I’m just taking the first and hoping for the best.)

That means there’s a brief window between the end of FixedUpdate and Unity’s physics pass during which a grounded actor might mistakenly believe it’s not on the ground, which is a bit of a shame, but there are very few good reasons for anything to be happening in that window.

Okay! Now we can do slopes.

Just kidding! First we have to do sliding.

When I first looked at this code, it didn’t apply gravity while the player was on the ground. I think I may have had some problems with detecting the ground as result, since the player was no longer pushing down against it? Either way, it seemed like a silly special case, so I made gravity always apply.

Lo! I was a fool. The player could no longer move.

Why? Because MovePosition does exactly what it promises. If the player collides with something, they’ll stop moving. Applying gravity means that the player is trying to move diagonally downwards into the ground, and so MovePosition stops them immediately.

Hence, sliding. I don’t want the player to actually try to move into the ground. I want them to move the unblocked part of that movement. For flat ground, that means the horizontal part, which is pretty much the same as discarding gravity. For sloped ground, it’s a bit more complicated!

Okay but actually it’s less complicated than you’d think. It can be done with some cross products fairly easily, but Unity makes it even easier with a couple casts. There’s a Vector3.ProjectOnPlane function that projects an arbitrary vector on a plane given by its normal — exactly the thing I want! So I apply that to the attempted movement before passing it along to MovePosition. I do the same thing with the current velocity, to prevent the player from accelerating infinitely downwards while standing on flat ground.

One other thing: I don’t actually use the detected ground normal for this. The player might be touching two ground surfaces at the same time, and I’d want to project on both of them. Instead, I use the player body’s GetContacts method, which returns contact points (and normals!) for everything the player is currently touching. I believe those contact points are tracked by the physics engine anyway, so asking for them doesn’t require any actual physics work.

(Looking at the code I have, I notice that I still only perform the slide for surfaces facing upwards — but I’d want to slide against sloped ceilings, too. Why did I do this? Maybe I should remove that.)

(Also, I’m pretty sure projecting a vector on a plane is non-commutative, which raises the question of which order the projections should happen in and what difference it makes. I don’t have a good answer.)

(I note that my LÖVE setup does something slightly different: it just tries whatever the movement ought to be, and if there’s a collision, then it projects — and tries again with the remaining movement. But I can’t ask Unity to do multiple moves in one physics update, alas.)

Okay! Now, slopes. But actually, with the above work done, slopes are most of the way there already.

One obvious problem is that the player tries to move horizontally even when on a slope, and the easy fix is to change their movement from speed * Vector2.right to speed * new Vector2(ground.y, -ground.x) while on the ground. That’s the ground normal rotated a quarter-turn clockwise, so for flat ground it still points to the right, and in general it points rightwards along the ground. (Note that it assumes the ground normal is a unit vector, but as far as I’m aware, that’s true for all the normals Unity gives you.)

Another issue is that if the player stands motionless on a slope, gravity will cause them to slowly slide down it — because the movement from gravity will be projected onto the slope, and unlike flat ground, the result is no longer zero. For conscious actors only, I counter this by adding the opposite factor to the player’s velocity as part of adding in their walking speed. This matches how the real world works, to some extent: when you’re standing on a hill, you’re exerting some small amount of effort just to stay in place.

(Note that slope resistance is not the same as friction. Okay, yes, in the real world, virtually all resistance to movement happens as a result of friction, but bracing yourself against the ground isn’t the same as being passively resisted.)

From here there are a lot of things you can do, depending on how you think slopes should be handled. You could make the player unable to walk up slopes that are too steep. You could make walking down a slope faster than walking up it. You could make jumping go along the ground normal, rather than straight up. You could raise the player’s max allowed speed while running downhill. Whatever you want, really. Armed with a normal and awareness of dot products, you can do whatever you want.

But first you might want to fix a few aggravating side effects.

Problem 3: Ground adherence

I don’t know if there’s a better name for this. I rarely even see anyone talk about it, which surprises me; it seems like it should be a very common problem.

The problem is: if the player runs up a slope which then abruptly changes to flat ground, their momentum will carry them into the air. For very fast players going off the top of very steep slopes, this makes sense, but it becomes visible even for relatively gentle slopes. It was a mild nightmare in the original release of our game Lunar Depot 38, which has very “rough” ground made up of lots of shallow slopes — so the player is very frequently slightly off the ground, which meant they couldn’t jump, for seemingly no reason. (I even had code to fix this, but I disabled it because of a silly visual side effect that I never got around to fixing.)

Anyway! The reason this is a problem is that game protagonists are generally not boxes sliding around — they have legs. We don’t go flying off the top of real-world hilltops because we put our foot down until it touches the ground.

Simulating this footfall is surprisingly fiddly to get right, especially with someone else’s physics engine. It’s made somewhat easier by Cast, which casts the entire hitbox — no matter what shape it is — in a particular direction, as if it had moved, and tells you all the hypothetical collisions in order.

So I cast the player in the direction of gravity by some distance. If the cast hits something solid with a ground-like collision normal, then the player must be close to the ground, and I move them down to touch it (and set that ground as the new ground normal).

There are some wrinkles.

Wrinkle 1: I only want to do this if the player is off the ground now, but was on the ground last frame, and is not deliberately moving upwards. That latter condition means I want to skip this logic if the player jumps, for example, but also if the player is thrust upwards by a spring or abducted by a UFO or whatever. As long as external code goes through some interface and doesn’t mess with the player’s velocity directly, that shouldn’t be too hard to track.

Wrinkle 2: When does this logic run? It needs to happen after the player moves, which means after a Unity physics pass… but there’s no callback for that point in time. I ended up running it at the beginning of FixedUpdate and the beginning of Update — since I definitely want to do it before rendering happens! That means it’ll sometimes happen twice between physics updates. (I could carefully juggle a flag to skip the second run, but I… didn’t do that. Yet?)

Wrinkle 3: I can’t move the player with MovePosition! Remember, MovePosition schedules a movement, it doesn’t actually perform one; that means if it’s called twice before the physics pass, the first call is effectively ignored. I can’t easily combine the drop with the player’s regular movement, for various fiddly reasons. I ended up doing it “by hand” using transform.Translate, which I think was the “old way” to do manual movement before MovePosition existed. I’m not totally sure if it activates triggers? For that matter, I’m not sure it even notices collisions — but since I did a full-body Cast, there shouldn’t be any anyway.

Wrinkle 4: What, exactly, is “some distance”? I’ve yet to find a satisfying answer for this. It seems like it ought to be based on the player’s current speed and the slope of the ground they’re moving along, but every time I’ve done that math, I’ve gotten totally ludicrous answers that sometimes exceed the size of a tile. But maybe that’s not wrong? Play around, I guess, and think about when the effect should “break” and the player should go flying off the top of a hill.

Wrinkle 5: It’s possible that the player will launch off a slope, hit something, and then be adhered to the ground where they wouldn’t have hit it. I don’t much like this edge case, but I don’t see a way around it either.

This problem is surprisingly awkward for how simple it sounds, and the solution isn’t entirely satisfying. Oh, well; the results are much nicer than the solution. As an added bonus, this also fixes occasional problems with running down a hill and becoming detached from the ground due to precision issues or whathaveyou.

Problem 4: One-way platforms

Ah, what a nightmare.

It took me ages just to figure out how to define one-way platforms. Only block when the player is moving downwards? Nope. Only block when the player is above the platform? Nuh-uh.

Well, okay, yes, those approaches might work for convex players and flat platforms. But what about… sloped, one-way platforms? There’s no reason you shouldn’t be able to have those. If Super Mario World can do it, surely Unity can do it almost 30 years later.

The trick is, again, to look at the collision normal. If it faces away from gravity, the player is hitting a ground-like surface, so the platform should block them. Otherwise (or if the player overlaps the platform), it shouldn’t.

Here’s the catch: Unity doesn’t have conditional collision. I can’t decide, on the fly, whether a collision should block or not. In fact, I think that by the time I get a callback like OnCollisionEnter2D, the physics pass is already over.

I could go the other way and use triggers (which are non-blocking), but then I have the opposite problem: I can’t stop the player on the fly. I could move them back to where they hit the trigger, but I envision all kinds of problems as a result. What if they were moving fast enough to activate something on the other side of the platform? What if something else moved to where I’m trying to shove them back to in the meantime? How does this interact with ground detection and listing contacts, which would rightly ignore a trigger as non-blocking?

I beat my head against this for a while, but the inability to respond to collision conditionally was a huge roadblock. It’s all the more infuriating a problem, because Unity ships with a one-way platform modifier thing. Unfortunately, it seems to have been implemented by someone who has never played a platformer. It’s literally one-way — the player is only allowed to move straight upwards through it, not in from the sides. It also tries to block the player if they’re moving downwards while inside the platform, which invokes clumsy rejection behavior. And this all seems to be built into the physics engine itself somehow, so I can’t simply copy whatever they did.

Eventually, I settled on the following. After calculating attempted movement (including sliding), just at the end of FixedUpdate, I do a Cast along the movement vector. I’m not thrilled about having to duplicate the physics engine’s own work, but I do filter to only things on a “one-way platform” physics layer, which should at least help. For each object the cast hits, I use Physics2D.IgnoreCollision to either ignore or un-ignore the collision between the player and the platform, depending on whether the collision was ground-like or not.

(A lot of people suggested turning off collision between layers, but that can’t possibly work — the player might be standing on one platform while inside another, and anyway, this should work for all actors!)

Again, wrinkles! But fewer this time. Actually, maybe just one: handling the case where the player already overlaps the platform. I can’t just check for that with e.g. OverlapCollider, because that doesn’t distinguish between overlapping and merely touching.

I came up with a fairly simple fix: if I was going to un-ignore the collision (i.e. make the platform block), and the cast distance is reported as zero (either already touching or overlapping), I simply do nothing instead. If I’m standing on the platform, I must have already set it blocking when I was approaching it from the top anyway; if I’m overlapping it, I must have already set it non-blocking to get here in the first place.

I can imagine a few cases where this might go wrong. Moving platforms, especially, are going to cause some interesting issues. But this is the best I can do with what I know, and it seems to work well enough so far.

Oh, and our player can deliberately drop down through platforms, which was easy enough to implement; I just decide the platform is always passable while some button is held down.

Problem 5: Pushers and carriers

I haven’t gotten to this yet! Oh boy, can’t wait. I implemented it in LÖVE, but my way was hilariously invasive; I’m hoping that having a physics engine that supports a handwaved “this pushes that” will help. Of course, you also have to worry about sticking to platforms, for which the recommended solution is apparently to parent the cargo to the platform, which sounds goofy to me? I guess I’ll find out when I throw myself at it later.

Overall result

I ended up with a fairly pleasant-feeling system that supports slopes and one-way platforms and whatnot, with all the same pieces as I came up with for LÖVE. The code somehow ended up as less of a mess, too, but it probably helps that I’ve been down this rabbit hole once before and kinda knew what I was aiming for this time.

Animation of a character running smoothly along the top of an irregular dinosaur skeleton

Sorry that I don’t have a big block of code for you to copy-paste into your project. I don’t think there are nearly enough narrative discussions of these fundamentals, though, so hopefully this is useful to someone. If not, well, look forward to ✨ my book, that I am writing ✨!

Introducing Gluon: a new library for machine learning from AWS and Microsoft

Post Syndicated from Ana Visneski original https://aws.amazon.com/blogs/aws/introducing-gluon-a-new-library-for-machine-learning-from-aws-and-microsoft/

Post by Dr. Matt Wood

Today, AWS and Microsoft announced Gluon, a new open source deep learning interface which allows developers to more easily and quickly build machine learning models, without compromising performance.

Gluon Logo

Gluon provides a clear, concise API for defining machine learning models using a collection of pre-built, optimized neural network components. Developers who are new to machine learning will find this interface more familiar to traditional code, since machine learning models can be defined and manipulated just like any other data structure. More seasoned data scientists and researchers will value the ability to build prototypes quickly and utilize dynamic neural network graphs for entirely new model architectures, all without sacrificing training speed.

Gluon is available in Apache MXNet today, a forthcoming Microsoft Cognitive Toolkit release, and in more frameworks over time.

Neural Networks vs Developers
Machine learning with neural networks (including ‘deep learning’) has three main components: data for training; a neural network model, and an algorithm which trains the neural network. You can think of the neural network in a similar way to a directed graph; it has a series of inputs (which represent the data), which connect to a series of outputs (the prediction), through a series of connected layers and weights. During training, the algorithm adjusts the weights in the network based on the error in the network output. This is the process by which the network learns; it is a memory and compute intensive process which can take days.

Deep learning frameworks such as Caffe2, Cognitive Toolkit, TensorFlow, and Apache MXNet are, in part, an answer to the question ‘how can we speed this process up? Just like query optimizers in databases, the more a training engine knows about the network and the algorithm, the more optimizations it can make to the training process (for example, it can infer what needs to be re-computed on the graph based on what else has changed, and skip the unaffected weights to speed things up). These frameworks also provide parallelization to distribute the computation process, and reduce the overall training time.

However, in order to achieve these optimizations, most frameworks require the developer to do some extra work: specifically, by providing a formal definition of the network graph, up-front, and then ‘freezing’ the graph, and just adjusting the weights.

The network definition, which can be large and complex with millions of connections, usually has to be constructed by hand. Not only are deep learning networks unwieldy, but they can be difficult to debug and it’s hard to re-use the code between projects.

The result of this complexity can be difficult for beginners and is a time-consuming task for more experienced researchers. At AWS, we’ve been experimenting with some ideas in MXNet around new, flexible, more approachable ways to define and train neural networks. Microsoft is also a contributor to the open source MXNet project, and were interested in some of these same ideas. Based on this, we got talking, and found we had a similar vision: to use these techniques to reduce the complexity of machine learning, making it accessible to more developers.

Enter Gluon: dynamic graphs, rapid iteration, scalable training
Gluon introduces four key innovations.

  1. Friendly API: Gluon networks can be defined using a simple, clear, concise code – this is easier for developers to learn, and much easier to understand than some of the more arcane and formal ways of defining networks and their associated weighted scoring functions.
  2. Dynamic networks: the network definition in Gluon is dynamic: it can bend and flex just like any other data structure. This is in contrast to the more common, formal, symbolic definition of a network which the deep learning framework has to effectively carve into stone in order to be able to effectively optimizing computation during training. Dynamic networks are easier to manage, and with Gluon, developers can easily ‘hybridize’ between these fast symbolic representations and the more friendly, dynamic ‘imperative’ definitions of the network and algorithms.
  3. The algorithm can define the network: the model and the training algorithm are brought much closer together. Instead of separate definitions, the algorithm can adjust the network dynamically during definition and training. Not only does this mean that developers can use standard programming loops, and conditionals to create these networks, but researchers can now define even more sophisticated algorithms and models which were not possible before. They are all easier to create, change, and debug.
  4. High performance operators for training: which makes it possible to have a friendly, concise API and dynamic graphs, without sacrificing training speed. This is a huge step forward in machine learning. Some frameworks bring a friendly API or dynamic graphs to deep learning, but these previous methods all incur a cost in terms of training speed. As with other areas of software, abstraction can slow down computation since it needs to be negotiated and interpreted at run time. Gluon can efficiently blend together a concise API with the formal definition under the hood, without the developer having to know about the specific details or to accommodate the compiler optimizations manually.

The team here at AWS, and our collaborators at Microsoft, couldn’t be more excited to bring these improvements to developers through Gluon. We’re already seeing quite a bit of excitement from developers and researchers alike.

Getting started with Gluon
Gluon is available today in Apache MXNet, with support coming for the Microsoft Cognitive Toolkit in a future release. We’re also publishing the front-end interface and the low-level API specifications so it can be included in other frameworks in the fullness of time.

You can get started with Gluon today. Fire up the AWS Deep Learning AMI with a single click and jump into one of 50 fully worked, notebook examples. If you’re a contributor to a machine learning framework, check out the interface specs on GitHub.

-Dr. Matt Wood

VHostScan – Virtual Host Scanner With Alias & Catch-All Detection

Post Syndicated from Darknet original https://www.darknet.org.uk/2017/10/vhostscan-virtual-host-scanner-with-alias-catch-all-detection/?utm_source=rss&utm_medium=social&utm_campaign=darknetfeed

VHostScan – Virtual Host Scanner With Alias & Catch-All Detection

VHostScan is a Python-based virtual host scanner that can be used with pivot tools, detect catch-all scenarios, aliases and dynamic default pages.

Features of VHostScan Virtual Host Scanner

  • Quickly highlight unique content in catch-all scenarios
  • Locate the outliers in catch-all scenarios where results have dynamic content on the page (such as the time)
  • Identify aliases by tweaking the unique depth of matches
  • Wordlist supports standard words and a variable to input a base hostname (for e.g.

Read the rest of VHostScan – Virtual Host Scanner With Alias & Catch-All Detection now! Only available at Darknet.

JavaScript got better while I wasn’t looking

Post Syndicated from Eevee original https://eev.ee/blog/2017/10/07/javascript-got-better-while-i-wasnt-looking/

IndustrialRobot has generously donated in order to inquire:

In the last few years there seems to have been a lot of activity with adding emojis to Unicode. Has there been an equal effort to add ‘real’ languages/glyph systems/etc?

And as always, if you don’t have anything to say on that topic, feel free to choose your own. :p

Yes.

I mean, each release of Unicode lists major new additions right at the top — Unicode 10, Unicode 9, Unicode 8, etc. They also keep fastidious notes, so you can also dig into how and why these new scripts came from, by reading e.g. the proposal for the addition of Zanabazar Square. I don’t think I have much to add here; I’m not a real linguist, I only play one on TV.

So with that out of the way, here’s something completely different!

A brief history of JavaScript

JavaScript was created in seven days, about eight thousand years ago. It was pretty rough, and it stayed rough for most of its life. But that was fine, because no one used it for anything besides having a trail of sparkles follow your mouse on their Xanga profile.

Then people discovered you could actually do a handful of useful things with JavaScript, and it saw a sharp uptick in usage. Alas, it stayed pretty rough. So we came up with polyfills and jQuerys and all kinds of miscellaneous things that tried to smooth over the rough parts, to varying degrees of success.

And… that’s it. That’s pretty much how things stayed for a while.


I have complicated feelings about JavaScript. I don’t hate it… but I certainly don’t enjoy it, either. It has some pretty neat ideas, like prototypical inheritance and “everything is a value”, but it buries them under a pile of annoying quirks and a woefully inadequate standard library. The DOM APIs don’t make things much better — they seem to be designed as though the target language were Java, rarely taking advantage of any interesting JavaScript features. And the places where the APIs overlap with the language are a hilarious mess: I have to check documentation every single time I use any API that returns a set of things, because there are at least three totally different conventions for handling that and I can’t keep them straight.

The funny thing is that I’ve been fairly happy to work with Lua, even though it shares most of the same obvious quirks as JavaScript. Both languages are weakly typed; both treat nonexistent variables and keys as simply false values, rather than errors; both have a single data structure that doubles as both a list and a map; both use 64-bit floating-point as their only numeric type (though Lua added integers very recently); both lack a standard object model; both have very tiny standard libraries. Hell, Lua doesn’t even have exceptions, not really — you have to fake them in much the same style as Perl.

And yet none of this bothers me nearly as much in Lua. The differences between the languages are very subtle, but combined they make a huge impact.

  • Lua has separate operators for addition and concatenation, so + is never ambiguous. It also has printf-style string formatting in the standard library.

  • Lua’s method calls are syntactic sugar: foo:bar() just means foo.bar(foo). Lua doesn’t even have a special this or self value; the invocant just becomes the first argument. In contrast, JavaScript invokes some hand-waved magic to set its contextual this variable, which has led to no end of confusion.

  • Lua has an iteration protocol, as well as built-in iterators for dealing with list-style or map-style data. JavaScript has a special dedicated Array type and clumsy built-in iteration syntax.

  • Lua has operator overloading and (surprisingly flexible) module importing.

  • Lua allows the keys of a map to be any value (though non-scalars are always compared by identity). JavaScript implicitly converts keys to strings — and since there’s no operator overloading, there’s no way to natively fix this.

These are fairly minor differences, in the grand scheme of language design. And almost every feature in Lua is implemented in a ridiculously simple way; in fact the entire language is described in complete detail in a single web page. So writing JavaScript is always frustrating for me: the language is so close to being much more ergonomic, and yet, it isn’t.

Or, so I thought. As it turns out, while I’ve been off doing other stuff for a few years, browser vendors have been implementing all this pie-in-the-sky stuff from “ES5” and “ES6”, whatever those are. People even upgrade their browsers now. Lo and behold, the last time I went to write JavaScript, I found out that a number of papercuts had actually been solved, and the solutions were sufficiently widely available that I could actually use them in web code.

The weird thing is that I do hear a lot about JavaScript, but the feature I’ve seen raved the most about by far is probably… built-in types for working with arrays of bytes? That’s cool and all, but not exactly the most pressing concern for me.

Anyway, if you also haven’t been keeping tabs on the world of JavaScript, here are some things we missed.

let

MDN docs — supported in Firefox 44, Chrome 41, IE 11, Safari 10

I’m pretty sure I first saw let over a decade ago. Firefox has supported it for ages, but you actually had to opt in by specifying JavaScript version 1.7. Remember JavaScript versions? You know, from back in the days when people actually suggested you write stuff like this:

1
<SCRIPT LANGUAGE="JavaScript1.2" TYPE="text/javascript">

Yikes.

Anyway, so, let declares a variable — but scoped to the immediately containing block, unlike var, which scopes to the innermost function. The trouble with var was that it was very easy to make misleading:

1
2
3
4
5
6
// foo exists here
while (true) {
    var foo = ...;
    ...
}
// foo exists here too

If you reused the same temporary variable name in a different block, or if you expected to be shadowing an outer foo, or if you were trying to do something with creating closures in a loop, this would cause you some trouble.

But no more, because let actually scopes the way it looks like it should, the way variable declarations do in C and friends. As an added bonus, if you refer to a variable declared with let outside of where it’s valid, you’ll get a ReferenceError instead of a silent undefined value. Hooray!

There’s one other interesting quirk to let that I can’t find explicitly documented. Consider:

1
2
3
4
5
6
7
let closures = [];
for (let i = 0; i < 4; i++) {
    closures.push(function() { console.log(i); });
}
for (let j = 0; j < closures.length; j++) {
    closures[j]();
}

If this code had used var i, then it would print 4 four times, because the function-scoped var i means each closure is sharing the same i, whose final value is 4. With let, the output is 0 1 2 3, as you might expect, because each run through the loop gets its own i.

But wait, hang on.

The semantics of a C-style for are that the first expression is only evaluated once, at the very beginning. So there’s only one let i. In fact, it makes no sense for each run through the loop to have a distinct i, because the whole idea of the loop is to modify i each time with i++.

I assume this is simply a special case, since it’s what everyone expects. We expect it so much that I can’t find anyone pointing out that the usual explanation for why it works makes no sense. It has the interesting side effect that for no longer de-sugars perfectly to a while, since this will print all 4s:

1
2
3
4
5
6
7
8
9
closures = [];
let i = 0;
while (i < 4) {
    closures.push(function() { console.log(i); });
    i++;
}
for (let j = 0; j < closures.length; j++) {
    closures[j]();
}

This isn’t a problem — I’m glad let works this way! — it just stands out to me as interesting. Lua doesn’t need a special case here, since it uses an iterator protocol that produces values rather than mutating a visible state variable, so there’s no problem with having the loop variable be truly distinct on each run through the loop.

Classes

MDN docs — supported in Firefox 45, Chrome 42, Safari 9, Edge 13

Prototypical inheritance is pretty cool. The way JavaScript presents it is a little bit opaque, unfortunately, which seems to confuse a lot of people. JavaScript gives you enough functionality to make it work, and even makes it sound like a first-class feature with a property outright called prototype… but to actually use it, you have to do a bunch of weird stuff that doesn’t much look like constructing an object or type.

The funny thing is, people with almost any background get along with Python just fine, and Python uses prototypical inheritance! Nobody ever seems to notice this, because Python tucks it neatly behind a class block that works enough like a Java-style class. (Python also handles inheritance without using the prototype, so it’s a little different… but I digress. Maybe in another post.)

The point is, there’s nothing fundamentally wrong with how JavaScript handles objects; the ergonomics are just terrible.

Lo! They finally added a class keyword. Or, rather, they finally made the class keyword do something; it’s been reserved this entire time.

 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;
    }
}

This is all just sugar for existing features: creating a Vector function to act as the constructor, assigning a function to Vector.prototype.dot, and whatever it is you do to make a property. (Oh, there are properties. I’ll get to that in a bit.)

The class block can be used as an expression, with or without a name. It also supports prototypical inheritance with an extends clause and has a super pseudo-value for superclass calls.

It’s a little weird that the inside of the class block has its own special syntax, with function omitted and whatnot, but honestly you’d have a hard time making a class block without special syntax.

One severe omission here is that you can’t declare values inside the block, i.e. you can’t just drop a bar = 3; in there if you want all your objects to share a default attribute. The workaround is to just do this.bar = 3; inside the constructor, but I find that unsatisfying, since it defeats half the point of using prototypes.

Properties

MDN docs — supported in Firefox 4, Chrome 5, IE 9, Safari 5.1

JavaScript historically didn’t have a way to intercept attribute access, which is a travesty. And by “intercept attribute access”, I mean that you couldn’t design a value foo such that evaluating foo.bar runs some code you wrote.

Exciting news: now it does. Or, rather, you can intercept specific attributes, like in the class example above. The above magnitude definition is equivalent to:

1
2
3
4
5
6
7
Object.defineProperty(Vector.prototype, 'magnitude', {
    configurable: true,
    enumerable: true,
    get: function() {
        return Math.sqrt(this.x * this.x + this.y * this.y);
    },
});

Beautiful.

And what even are these configurable and enumerable things? It seems that every single key on every single object now has its own set of three Boolean twiddles:

  • configurable means the property itself can be reconfigured with another call to Object.defineProperty.
  • enumerable means the property appears in for..in or Object.keys().
  • writable means the property value can be changed, which only applies to properties with real values rather than accessor functions.

The incredibly wild thing is that for properties defined by Object.defineProperty, configurable and enumerable default to false, meaning that by default accessor properties are immutable and invisible. Super weird.

Nice to have, though. And luckily, it turns out the same syntax as in class also works in object literals.

1
2
3
4
5
6
Vector.prototype = {
    get magnitude() {
        return Math.sqrt(this.x * this.x + this.y * this.y);
    },
    ...
};

Alas, I’m not aware of a way to intercept arbitrary attribute access.

Another feature along the same lines is Object.seal(), which marks all of an object’s properties as non-configurable and prevents any new properties from being added to the object. The object is still mutable, but its “shape” can’t be changed. And of course you can just make the object completely immutable if you want, via setting all its properties non-writable, or just using Object.freeze().

I have mixed feelings about the ability to irrevocably change something about a dynamic runtime. It would certainly solve some gripes of former Haskell-minded colleagues, and I don’t have any compelling argument against it, but it feels like it violates some unwritten contract about dynamic languages — surely any structural change made by user code should also be able to be undone by user code?

Slurpy arguments

MDN docs — supported in Firefox 15, Chrome 47, Edge 12, Safari 10

Officially this feature is called “rest parameters”, but that’s a terrible name, no one cares about “arguments” vs “parameters”, and “slurpy” is a good word. Bless you, Perl.

1
2
3
function foo(a, b, ...args) {
    // ...
}

Now you can call foo with as many arguments as you want, and every argument after the second will be collected in args as a regular array.

You can also do the reverse with the spread operator:

1
2
3
4
5
let args = [];
args.push(1);
args.push(2);
args.push(3);
foo(...args);

It even works in array literals, even multiple times:

1
2
let args2 = [...args, ...args];
console.log(args2);  // [1, 2, 3, 1, 2, 3]

Apparently there’s also a proposal for allowing the same thing with objects inside object literals.

Default arguments

MDN docs — supported in Firefox 15, Chrome 49, Edge 14, Safari 10

Yes, arguments can have defaults now. It’s more like Sass than Python — default expressions are evaluated once per call, and later default expressions can refer to earlier arguments. I don’t know how I feel about that but whatever.

1
2
3
function foo(n = 1, m = n + 1, list = []) {
    ...
}

Also, unlike Python, you can have an argument with a default and follow it with an argument without a default, since the default default (!) is and always has been defined as undefined. Er, let me just write it out.

1
2
3
function bar(a = 5, b) {
    ...
}

Arrow functions

MDN docs — supported in Firefox 22, Chrome 45, Edge 12, Safari 10

Perhaps the most humble improvement is the arrow function. It’s a slightly shorter way to write an anonymous function.

1
2
3
(a, b, c) => { ... }
a => { ... }
() => { ... }

An arrow function does not set this or some other magical values, so you can safely use an arrow function as a quick closure inside a method without having to rebind this. Hooray!

Otherwise, arrow functions act pretty much like regular functions; you can even use all the features of regular function signatures.

Arrow functions are particularly nice in combination with all the combinator-style array functions that were added a while ago, like Array.forEach.

1
2
3
[7, 8, 9].forEach(value => {
    console.log(value);
});

Symbol

MDN docs — supported in Firefox 36, Chrome 38, Edge 12, Safari 9

This isn’t quite what I’d call an exciting feature, but it’s necessary for explaining the next one. It’s actually… extremely weird.

symbol is a new kind of primitive (like number and string), not an object (like, er, Number and String). A symbol is created with Symbol('foo'). No, not new Symbol('foo'); that throws a TypeError, for, uh, some reason.

The only point of a symbol is as a unique key. You see, symbols have one very special property: they can be used as object keys, and will not be stringified. Remember, only strings can be keys in JavaScript — even the indices of an array are, semantically speaking, still strings. Symbols are a new exception to this rule.

Also, like other objects, two symbols don’t compare equal to each other: Symbol('foo') != Symbol('foo').

The result is that symbols solve one of the problems that plauges most object systems, something I’ve talked about before: interfaces. Since an interface might be implemented by any arbitrary type, and any arbitrary type might want to implement any number of arbitrary interfaces, all the method names on an interface are effectively part of a single global namespace.

I think I need to take a moment to justify that. If you have IFoo and IBar, both with a method called method, and you want to implement both on the same type… you have a problem. Because most object systems consider “interface” to mean “I have a method called method, with no way to say which interface’s method you mean. This is a hard problem to avoid, because IFoo and IBar might not even come from the same library. Occasionally languages offer a clumsy way to “rename” one method or the other, but the most common approach seems to be for interface designers to avoid names that sound “too common”. You end up with redundant mouthfuls like IFoo.foo_method.

This incredibly sucks, and the only languages I’m aware of that avoid the problem are the ML family and Rust. In Rust, you define all the methods for a particular trait (interface) in a separate block, away from the type’s “own” methods. It’s pretty slick. You can still do obj.method(), and as long as there’s only one method among all the available traits, you’ll get that one. If not, there’s syntax for explicitly saying which trait you mean, which I can’t remember because I’ve never had to use it.

Symbols are JavaScript’s answer to this problem. If you want to define some interface, you can name its methods with symbols, which are guaranteed to be unique. You just have to make sure you keep the symbol around somewhere accessible so other people can actually use it. (Or… not?)

The interesting thing is that JavaScript now has several of its own symbols built in, allowing user objects to implement features that were previously reserved for built-in types. For example, you can use the Symbol.hasInstance symbol — which is simply where the language is storing an existing symbol and is not the same as Symbol('hasInstance')! — to override instanceof:

1
2
3
4
5
6
7
8
// oh my god don't do this though
class EvenNumber {
    static [Symbol.hasInstance](obj) {
        return obj % 2 == 0;
    }
}
console.log(2 instanceof EvenNumber);  // true
console.log(3 instanceof EvenNumber);  // false

Oh, and those brackets around Symbol.hasInstance are a sort of reverse-quoting — they indicate an expression to use where the language would normally expect a literal identifier. I think they work as object keys, too, and maybe some other places.

The equivalent in Python is to implement a method called __instancecheck__, a name which is not special in any way except that Python has reserved all method names of the form __foo__. That’s great for Python, but doesn’t really help user code. JavaScript has actually outclassed (ho ho) Python here.

Of course, obj[BobNamespace.some_method]() is not the prettiest way to call an interface method, so it’s not perfect. I imagine this would be best implemented in user code by exposing a polymorphic function, similar to how Python’s len(obj) pretty much just calls obj.__len__().

I only bring this up because it’s the plumbing behind one of the most incredible things in JavaScript that I didn’t even know about until I started writing this post. I’m so excited oh my gosh. Are you ready? It’s:

Iteration protocol

MDN docs — supported in Firefox 27, Chrome 39, Safari 10; still experimental in Edge

Yes! Amazing! JavaScript has first-class support for iteration! I can’t even believe this.

It works pretty much how you’d expect, or at least, how I’d expect. You give your object a method called Symbol.iterator, and that returns an iterator.

What’s an iterator? It’s an object with a next() method that returns the next value and whether the iterator is exhausted.

Wait, wait, wait a second. Hang on. The method is called next? Really? You didn’t go for Symbol.next? Python 2 did exactly the same thing, then realized its mistake and changed it to __next__ in Python 3. Why did you do this?

Well, anyway. My go-to test of an iterator protocol is how hard it is to write an equivalent to Python’s enumerate(), which takes a list and iterates over its values and their indices. In Python it looks like this:

1
2
3
4
5
for i, value in enumerate(['one', 'two', 'three']):
    print(i, value)
# 0 one
# 1 two
# 2 three

It’s super nice to have, and I’m always amazed when languages with “strong” “support” for iteration don’t have it. Like, C# doesn’t. So if you want to iterate over a list but also need indices, you need to fall back to a C-style for loop. And if you want to iterate over a lazy or arbitrary iterable but also need indices, you need to track it yourself with a counter. Ridiculous.

Here’s my attempt at building it in JavaScript.

 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
function enumerate(iterable) {
    // Return a new iter*able* object with a Symbol.iterator method that
    // returns an iterator.
    return {
        [Symbol.iterator]: function() {
            let iterator = iterable[Symbol.iterator]();
            let i = 0;

            return {
                next: function() {
                    let nextval = iterator.next();
                    if (! nextval.done) {
                        nextval.value = [i, nextval.value];
                        i++;
                    }
                    return nextval;
                },
            };
        },
    };
}
for (let [i, value] of enumerate(['one', 'two', 'three'])) {
    console.log(i, value);
}
// 0 one
// 1 two
// 2 three

Incidentally, for..of (which iterates over a sequence, unlike for..in which iterates over keys — obviously) is finally supported in Edge 12. Hallelujah.

Oh, and let [i, value] is destructuring assignment, which is also a thing now and works with objects as well. You can even use the splat operator with it! Like Python! (And you can use it in function signatures! Like Python! Wait, no, Python decided that was terrible and removed it in 3…)

1
let [x, y, ...others] = ['apple', 'orange', 'cherry', 'banana'];

It’s a Halloween miracle. 🎃

Generators

MDN docs — supported in Firefox 26, Chrome 39, Edge 13, Safari 10

That’s right, JavaScript has goddamn generators now. It’s basically just copying Python and adding a lot of superfluous punctuation everywhere. Not that I’m complaining.

Also, generators are themselves iterable, so I’m going to cut to the chase and rewrite my enumerate() with a generator.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
function enumerate(iterable) {
    return {
        [Symbol.iterator]: function*() {
            let i = 0;
            for (let value of iterable) {
                yield [i, value];
                i++;
            }
        },
    };
}
for (let [i, value] of enumerate(['one', 'two', 'three'])) {
    console.log(i, value);
}
// 0 one
// 1 two
// 2 three

Amazing. function* is a pretty strange choice of syntax, but whatever? I guess it also lets them make yield only act as a keyword inside a generator, for ultimate backwards compatibility.

JavaScript generators support everything Python generators do: yield* yields every item from a subsequence, like Python’s yield from; generators can return final values; you can pass values back into the generator if you iterate it by hand. No, really, I wasn’t kidding, it’s basically just copying Python. It’s great. You could now built asyncio in JavaScript!

In fact, they did that! JavaScript now has async and await. An async function returns a Promise, which is also a built-in type now. Amazing.

Sets and maps

MDN docs for MapMDN docs for Set — supported in Firefox 13, Chrome 38, IE 11, Safari 7.1

I did not save the best for last. This is much less exciting than generators. But still exciting.

The only data structure in JavaScript is the object, a map where the strings are keys. (Or now, also symbols, I guess.) That means you can’t readily use custom values as keys, nor simulate a set of arbitrary objects. And you have to worry about people mucking with Object.prototype, yikes.

But now, there’s Map and Set! Wow.

Unfortunately, because JavaScript, Map couldn’t use the indexing operators without losing the ability to have methods, so you have to use a boring old method-based API. But Map has convenient methods that plain objects don’t, like entries() to iterate over pairs of keys and values. In fact, you can use a map with for..of to get key/value pairs. So that’s nice.

Perhaps more interesting, there’s also now a WeakMap and WeakSet, where the keys are weak references. I don’t think JavaScript had any way to do weak references before this, so that’s pretty slick. There’s no obvious way to hold a weak value, but I guess you could substitute a WeakSet with only one item.

Template literals

MDN docs — supported in Firefox 34, Chrome 41, Edge 12, Safari 9

Template literals are JavaScript’s answer to string interpolation, which has historically been a huge pain in the ass because it doesn’t even have string formatting in the standard library.

They’re just strings delimited by backticks instead of quotes. They can span multiple lines and contain expressions.

1
2
console.log(`one plus
two is ${1 + 2}`);

Someone decided it would be a good idea to allow nesting more sets of backticks inside a ${} expression, so, good luck to syntax highlighters.

However, someone also had the most incredible idea ever, which was to add syntax allowing user code to do the interpolation — so you can do custom escaping, when absolutely necessary, which is virtually never, because “escaping” means you’re building a structured format by slopping strings together willy-nilly instead of using some API that works with the structure.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
// OF COURSE, YOU SHOULDN'T BE DOING THIS ANYWAY; YOU SHOULD BUILD HTML WITH
// THE DOM API AND USE .textContent FOR LITERAL TEXT.  BUT AS AN EXAMPLE:
function html(literals, ...values) {
    let ret = [];
    literals.forEach((literal, i) => {
        if (i > 0) {
            // Is there seriously still not a built-in function for doing this?
            // Well, probably because you SHOULDN'T BE DOING IT
            ret.push(values[i - 1]
                .replace(/&/g, '&amp;')
                .replace(/</g, '&lt;')
                .replace(/>/g, '&gt;')
                .replace(/"/g, '&quot;')
                .replace(/'/g, '&apos;'));
        }
        ret.push(literal);
    });
    return ret.join('');
}
let username = 'Bob<script>';
let result = html`<b>Hello, ${username}!</b>`;
console.log(result);
// <b>Hello, Bob&lt;script&gt;!</b>

It’s a shame this feature is in JavaScript, the language where you are least likely to need it.

Trailing commas

Remember how you couldn’t do this for ages, because ass-old IE considered it a syntax error and would reject the entire script?

1
2
3
4
5
{
    a: 'one',
    b: 'two',
    c: 'three',  // <- THIS GUY RIGHT HERE
}

Well now it’s part of the goddamn spec and if there’s anything in this post you can rely on, it’s this. In fact you can use AS MANY GODDAMN TRAILING COMMAS AS YOU WANT. But only in arrays.

1
[1, 2, 3,,,,,,,,,,,,,,,,,,,,,,,,,]

Apparently that has the bizarre side effect of reserving extra space at the end of the array, without putting values there.

And more, probably

Like strict mode, which makes a few silent “errors” be actual errors, forces you to declare variables (no implicit globals!), and forbids the completely bozotic with block.

Or String.trim(), which trims whitespace off of strings.

Or… Math.sign()? That’s new? Seriously? Well, okay.

Or the Proxy type, which lets you customize indexing and assignment and calling. Oh. I guess that is possible, though this is a pretty weird way to do it; why not just use symbol-named methods?

You can write Unicode escapes for astral plane characters in strings (or identifiers!), as \u{XXXXXXXX}.

There’s a const now? I extremely don’t care, just name it in all caps and don’t reassign it, come on.

There’s also a mountain of other minor things, which you can peruse at your leisure via MDN or the ECMAScript compatibility tables (note the links at the top, too).

That’s all I’ve got. I still wouldn’t say I’m a big fan of JavaScript, but it’s definitely making an effort to clean up some goofy inconsistencies and solve common problems. I think I could even write some without yelling on Twitter about it now.

On the other hand, if you’re still stuck supporting IE 10 for some reason… well, er, my condolences.

systemd 235 released

Post Syndicated from corbet original https://lwn.net/Articles/735691/rss

Version 235 of the systemd service manager is out; it includes a long list
of new features. See this blog
post
for a description of the dynamic user feature in particular.
One major benefit of dynamic user IDs is that running a
privilege-separated service leaves no artifacts in the system. A system
user is allocated and made use of, but it is discarded automatically in a
safe and secure way after use, in a fashion that is safe for later
recycling. Thus, quickly invoking a short-lived service for processing some
job can be protected properly through a user ID without having to
pre-allocate it and without this draining the available UID pool any longer
than necessary.

Dynamic Users with systemd

Post Syndicated from Lennart Poettering original http://0pointer.net/blog/dynamic-users-with-systemd.html

TL;DR: you may now configure systemd to dynamically allocate a UNIX
user ID for service processes when it starts them and release it when
it stops them. It’s pretty secure, mixes well with transient services,
socket activated services and service templating.

Today we released systemd
235
. Among
other improvements this greatly extends the dynamic user logic of
systemd. Dynamic users are a powerful but little known concept,
supported in its basic form since systemd 232. With this blog story I
hope to make it a bit better known.

The UNIX user concept is the most basic and well-understood security
concept in POSIX operating systems. It is UNIX/POSIX’ primary security
concept, the one everybody can agree on, and most security concepts
that came after it (such as process capabilities, SELinux and other
MACs, user name-spaces, …) in some form or another build on it, extend
it or at least interface with it. If you build a Linux kernel with all
security features turned off, the user concept is pretty much the one
you’ll still retain.

Originally, the user concept was introduced to make multi-user systems
a reality, i.e. systems enabling multiple human users to share the
same system at the same time, cleanly separating their resources and
protecting them from each other. The majority of today’s UNIX systems
don’t really use the user concept like that anymore though. Most of
today’s systems probably have only one actual human user (or even
less!), but their user databases (/etc/passwd) list a good number
more entries than that. Today, the majority of UNIX users in most
environments are system users, i.e. users that are not the technical
representation of a human sitting in front of a PC anymore, but the
security identity a system service — an executable program — runs
as. Event though traditional, simultaneous multi-user systems slowly
became less relevant, their ground-breaking basic concept became the
cornerstone of UNIX security. The OS is nowadays partitioned into
isolated services — and each service runs as its own system user, and
thus within its own, minimal security context.

The people behind the Android OS realized the relevance of the UNIX
user concept as the primary security concept on UNIX, and took its use
even further: on Android not only system services take benefit of the
UNIX user concept, but each UI app gets its own, individual user
identity too — thus neatly separating app resources from each other,
and protecting app processes from each other, too.

Back in the more traditional Linux world things are a bit less
advanced in this area. Even though users are the quintessential UNIX
security concept, allocation and management of system users is still a
pretty limited, raw and static affair. In most cases, RPM or DEB
package installation scripts allocate a fixed number of (usually one)
system users when you install the package of a service that wants to
take benefit of the user concept, and from that point on the system
user remains allocated on the system and is never deallocated again,
even if the package is later removed again. Most Linux distributions
limit the number of system users to 1000 (which isn’t particularly a
lot). Allocating a system user is hence expensive: the number of
available users is limited, and there’s no defined way to dispose of
them after use. If you make use of system users too liberally, you are
very likely to run out of them sooner rather than later.

You may wonder why system users are generally not deallocated when the
package that registered them is uninstalled from a system (at least on
most distributions). The reason for that is one relevant property of
the user concept (you might even want to call this a design flaw):
user IDs are sticky to files (and other objects such as IPC
objects). If a service running as a specific system user creates a
file at some location, and is then terminated and its package and user
removed, then the created file still belongs to the numeric ID (“UID”)
the system user originally got assigned. When the next system user is
allocated and — due to ID recycling — happens to get assigned the same
numeric ID, then it will also gain access to the file, and that’s
generally considered a problem, given that the file belonged to a
potentially very different service once upon a time, and likely should
not be readable or changeable by anything coming after
it. Distributions hence tend to avoid UID recycling which means system
users remain registered forever on a system after they have been
allocated once.

The above is a description of the status quo ante. Let’s now focus on
what systemd’s dynamic user concept brings to the table, to improve
the situation.

Introducing Dynamic Users

With systemd dynamic users we hope to make make it easier and cheaper
to allocate system users on-the-fly, thus substantially increasing the
possible uses of this core UNIX security concept.

If you write a systemd service unit file, you may enable the dynamic
user logic for it by setting the
DynamicUser=
option in its [Service] section to yes. If you do a system user is
dynamically allocated the instant the service binary is invoked, and
released again when the service terminates. The user is automatically
allocated from the UID range 61184–65519, by looking for a so far
unused UID.

Now you may wonder, how does this concept deal with the sticky user
issue discussed above? In order to counter the problem, two strategies
easily come to mind:

  1. Prohibit the service from creating any files/directories or IPC objects

  2. Automatically removing the files/directories or IPC objects the
    service created when it shuts down.

In systemd we implemented both strategies, but for different parts of
the execution environment. Specifically:

  1. Setting DynamicUser=yes implies
    ProtectSystem=strict
    and
    ProtectHome=read-only. These
    sand-boxing options turn off write access to pretty much the whole OS
    directory tree, with a few relevant exceptions, such as the API file
    systems /proc, /sys and so on, as well as /tmp and
    /var/tmp. (BTW: setting these two options on your regular services
    that do not use DynamicUser= is a good idea too, as it drastically
    reduces the exposure of the system to exploited services.)

  2. Setting DynamicUser=yes implies
    PrivateTmp=yes. This
    option sets up /tmp and /var/tmp for the service in a way that it
    gets its own, disconnected version of these directories, that are not
    shared by other services, and whose life-cycle is bound to the
    service’s own life-cycle. Thus if the service goes down, the user is
    removed and all its temporary files and directories with it. (BTW: as
    above, consider setting this option for your regular services that do
    not use DynamicUser= too, it’s a great way to lock things down
    security-wise.)

  3. Setting DynamicUser=yes implies
    RemoveIPC=yes. This
    option ensures that when the service goes down all SysV and POSIX IPC
    objects (shared memory, message queues, semaphores) owned by the
    service’s user are removed. Thus, the life-cycle of the IPC objects is
    bound to the life-cycle of the dynamic user and service, too. (BTW:
    yes, here too, consider using this in your regular services, too!)

With these four settings in effect, services with dynamic users are
nicely sand-boxed. They cannot create files or directories, except in
/tmp and /var/tmp, where they will be removed automatically when
the service shuts down, as will any IPC objects created. Sticky
ownership of files/directories and IPC objects is hence dealt with
effectively.

The
RuntimeDirectory=
option may be used to open up a bit the sandbox to external
programs. If you set it to a directory name of your choice, it will be
created below /run when the service is started, and removed in its
entirety when it is terminated. The ownership of the directory is
assigned to the service’s dynamic user. This way, a dynamic user
service can expose API interfaces (AF_UNIX sockets, …) to other
services at a well-defined place and again bind the life-cycle of it to
the service’s own run-time. Example: set RuntimeDirectory=foobar in
your service, and watch how a directory /run/foobar appears at the
moment you start the service, and disappears the moment you stop
it again. (BTW: Much like the other settings discussed above,
RuntimeDirectory= may be used outside of the DynamicUser= context
too, and is a nice way to run any service with a properly owned,
life-cycle-managed run-time directory.)

Persistent Data

Of course, a service running in such an environment (although already
very useful for many cases!), has a major limitation: it cannot leave
persistent data around it can reuse on a later run. As pretty much the
whole OS directory tree is read-only to it, there’s simply no place it
could put the data that survives from one service invocation to the
next.

With systemd 235 this limitation is removed: there are now three new
settings:
StateDirectory=,
LogsDirectory= and CacheDirectory=. In many ways they operate like
RuntimeDirectory=, but create sub-directories below /var/lib,
/var/log and /var/cache, respectively. There’s one major
difference beyond that however: directories created that way are
persistent, they will survive the run-time cycle of a service, and
thus may be used to store data that is supposed to stay around between
invocations of the service.

Of course, the obvious question to ask now is: how do these three
settings deal with the sticky file ownership problem?

For that we lifted a concept from container managers. Container
managers have a very similar problem: each container and the host
typically end up using a very similar set of numeric UIDs, and unless
user name-spacing is deployed this means that host users might be able
to access the data of specific containers that also have a user by the
same numeric UID assigned, even though it actually refers to a very
different identity in a different context. (Actually, it’s even worse
than just getting access, due to the existence of setuid file bits,
access might translate to privilege elevation.) The way container
managers protect the container images from the host (and from each
other to some level) is by placing the container trees below a
boundary directory, with very restrictive access modes and ownership
(0700 and root:root or so). A host user hence cannot take advantage
of the files/directories of a container user of the same UID inside of
a local container tree, simply because the boundary directory makes it
impossible to even reference files in it. After all on UNIX, in order
to get access to a specific path you need access to every single
component of it.

How is that applied to dynamic user services? Let’s say
StateDirectory=foobar is set for a service that has DynamicUser=
turned off. The instant the service is started, /var/lib/foobar is
created as state directory, owned by the service’s user and remains in
existence when the service is stopped. If the same service now is run
with DynamicUser= turned on, the implementation is slightly
altered. Instead of a directory /var/lib/foobar a symbolic link by
the same path is created (owned by root), pointing to
/var/lib/private/foobar (the latter being owned by the service’s
dynamic user). The /var/lib/private directory is created as boundary
directory: it’s owned by root:root, and has a restrictive access
mode of 0700. Both the symlink and the service’s state directory will
survive the service’s life-cycle, but the state directory will remain,
and continues to be owned by the now disposed dynamic UID — however it
is protected from other host users (and other services which might get
the same dynamic UID assigned due to UID recycling) by the boundary
directory.

The obvious question to ask now is: but if the boundary directory
prohibits access to the directory from unprivileged processes, how can
the service itself which runs under its own dynamic UID access it
anyway? This is achieved by invoking the service process in a slightly
modified mount name-space: it will see most of the file hierarchy the
same way as everything else on the system (modulo /tmp and
/var/tmp as mentioned above), except for /var/lib/private, which
is over-mounted with a read-only tmpfs file system instance, with a
slightly more liberal access mode permitting the service read
access. Inside of this tmpfs file system instance another mount is
placed: a bind mount to the host’s real /var/lib/private/foobar
directory, onto the same name. Putting this together these means that
superficially everything looks the same and is available at the same
place on the host and from inside the service, but two important
changes have been made: the /var/lib/private boundary directory lost
its restrictive character inside the service, and has been emptied of
the state directories of any other service, thus making the protection
complete. Note that the symlink /var/lib/foobar hides the fact that
the boundary directory is used (making it little more than an
implementation detail), as the directory is available this way under
the same name as it would be if DynamicUser= was not used. Long
story short: for the daemon and from the view from the host the
indirection through /var/lib/private is mostly transparent.

This logic of course raises another question: what happens to the
state directory if a dynamic user service is started with a state
directory configured, gets UID X assigned on this first invocation,
then terminates and is restarted and now gets UID Y assigned on the
second invocation, with X ≠ Y? On the second invocation the directory
— and all the files and directories below it — will still be owned by
the original UID X so how could the second instance running as Y
access it? Our way out is simple: systemd will recursively change the
ownership of the directory and everything contained within it to UID Y
before invoking the service’s executable.

Of course, such recursive ownership changing (chown()ing) of whole
directory trees can become expensive (though according to my
experiences, IRL and for most services it’s much cheaper than you
might think), hence in order to optimize behavior in this regard, the
allocation of dynamic UIDs has been tweaked in two ways to avoid the
necessity to do this expensive operation in most cases: firstly, when
a dynamic UID is allocated for a service an allocation loop is
employed that starts out with a UID hashed from the service’s
name. This means a service by the same name is likely to always use
the same numeric UID. That means that a stable service name translates
into a stable dynamic UID, and that means recursive file ownership
adjustments can be skipped (of course, after validation). Secondly, if
the configured state directory already exists, and is owned by a
suitable currently unused dynamic UID, it’s preferably used above
everything else, thus maximizing the chance we can avoid the
chown()ing. (That all said, ultimately we have to face it, the
currently available UID space of 4K+ is very small still, and
conflicts are pretty likely sooner or later, thus a chown()ing has to
be expected every now and then when this feature is used extensively).

Note that CacheDirectory= and LogsDirectory= work very similar to
StateDirectory=. The only difference is that they manage directories
below the /var/cache and /var/logs directories, and their boundary
directory hence is /var/cache/private and /var/log/private,
respectively.

Examples

So, after all this introduction, let’s have a look how this all can be
put together. Here’s a trivial example:

# cat > /etc/systemd/system/dynamic-user-test.service <<EOF
[Service]
ExecStart=/usr/bin/sleep 4711
DynamicUser=yes
EOF
# systemctl daemon-reload
# systemctl start dynamic-user-test
# systemctl status dynamic-user-test
● dynamic-user-test.service
   Loaded: loaded (/etc/systemd/system/dynamic-user-test.service; static; vendor preset: disabled)
   Active: active (running) since Fri 2017-10-06 13:12:25 CEST; 3s ago
 Main PID: 2967 (sleep)
    Tasks: 1 (limit: 4915)
   CGroup: /system.slice/dynamic-user-test.service
           └─2967 /usr/bin/sleep 4711

Okt 06 13:12:25 sigma systemd[1]: Started dynamic-user-test.service.
# ps -e -o pid,comm,user | grep 2967
 2967 sleep           dynamic-user-test
# id dynamic-user-test
uid=64642(dynamic-user-test) gid=64642(dynamic-user-test) groups=64642(dynamic-user-test)
# systemctl stop dynamic-user-test
# id dynamic-user-test
id: ‘dynamic-user-test’: no such user

In this example, we create a unit file with DynamicUser= turned on,
start it, check if it’s running correctly, have a look at the service
process’ user (which is named like the service; systemd does this
automatically if the service name is suitable as user name, and you
didn’t configure any user name to use explicitly), stop the service
and verify that the user ceased to exist too.

That’s already pretty cool. Let’s step it up a notch, by doing the
same in an interactive transient service (for those who don’t know
systemd well: a transient service is a service that is defined and
started dynamically at run-time, for example via the systemd-run
command from the shell. Think: run a service without having to write a
unit file first):

# systemd-run --pty --property=DynamicUser=yes --property=StateDirectory=wuff /bin/sh
Running as unit: run-u15750.service
Press ^] three times within 1s to disconnect TTY.
sh-4.4$ id
uid=63122(run-u15750) gid=63122(run-u15750) groups=63122(run-u15750) context=system_u:system_r:initrc_t:s0
sh-4.4$ ls -al /var/lib/private/
total 0
drwxr-xr-x. 3 root       root        60  6. Okt 13:21 .
drwxr-xr-x. 1 root       root       852  6. Okt 13:21 ..
drwxr-xr-x. 1 run-u15750 run-u15750   8  6. Okt 13:22 wuff
sh-4.4$ ls -ld /var/lib/wuff
lrwxrwxrwx. 1 root root 12  6. Okt 13:21 /var/lib/wuff -> private/wuff
sh-4.4$ ls -ld /var/lib/wuff/
drwxr-xr-x. 1 run-u15750 run-u15750 0  6. Okt 13:21 /var/lib/wuff/
sh-4.4$ echo hello > /var/lib/wuff/test
sh-4.4$ exit
exit
# id run-u15750
id: ‘run-u15750’: no such user
# ls -al /var/lib/private
total 0
drwx------. 1 root  root   66  6. Okt 13:21 .
drwxr-xr-x. 1 root  root  852  6. Okt 13:21 ..
drwxr-xr-x. 1 63122 63122   8  6. Okt 13:22 wuff
# ls -ld /var/lib/wuff
lrwxrwxrwx. 1 root root 12  6. Okt 13:21 /var/lib/wuff -> private/wuff
# ls -ld /var/lib/wuff/
drwxr-xr-x. 1 63122 63122 8  6. Okt 13:22 /var/lib/wuff/
# cat /var/lib/wuff/test
hello

The above invokes an interactive shell as transient service
run-u15750.service (systemd-run picked that name automatically,
since we didn’t specify anything explicitly) with a dynamic user whose
name is derived automatically from the service name. Because
StateDirectory=wuff is used, a persistent state directory for the
service is made available as /var/lib/wuff. In the interactive shell
running inside the service, the ls commands show the
/var/lib/private boundary directory and its contents, as well as the
symlink that is placed for the service. Finally, before exiting the
shell, a file is created in the state directory. Back in the original
command shell we check if the user is still allocated: it is not, of
course, since the service ceased to exist when we exited the shell and
with it the dynamic user associated with it. From the host we check
the state directory of the service, with similar commands as we did
from inside of it. We see that things are set up pretty much the same
way in both cases, except for two things: first of all the user/group
of the files is now shown as raw numeric UIDs instead of the
user/group names derived from the unit name. That’s because the user
ceased to exist at this point, and “ls” shows the raw UID for files
owned by users that don’t exist. Secondly, the access mode of the
boundary directory is different: when we look at it from outside of
the service it is not readable by anyone but root, when we looked from
inside we saw it it being world readable.

Now, let’s see how things look if we start another transient service,
reusing the state directory from the first invocation:

# systemd-run --pty --property=DynamicUser=yes --property=StateDirectory=wuff /bin/sh
Running as unit: run-u16087.service
Press ^] three times within 1s to disconnect TTY.
sh-4.4$ cat /var/lib/wuff/test
hello
sh-4.4$ ls -al /var/lib/wuff/
total 4
drwxr-xr-x. 1 run-u16087 run-u16087  8  6. Okt 13:22 .
drwxr-xr-x. 3 root       root       60  6. Okt 15:42 ..
-rw-r--r--. 1 run-u16087 run-u16087  6  6. Okt 13:22 test
sh-4.4$ id
uid=63122(run-u16087) gid=63122(run-u16087) groups=63122(run-u16087) context=system_u:system_r:initrc_t:s0
sh-4.4$ exit
exit

Here, systemd-run picked a different auto-generated unit name, but
the used dynamic UID is still the same, as it was read from the
pre-existing state directory, and was otherwise unused. As we can see
the test file we generated earlier is accessible and still contains
the data we left in there. Do note that the user name is different
this time (as it is derived from the unit name, which is different),
but the UID it is assigned to is the same one as on the first
invocation. We can thus see that the mentioned optimization of the UID
allocation logic (i.e. that we start the allocation loop from the UID
owner of any existing state directory) took effect, so that no
recursive chown()ing was required.

And that’s the end of our example, which hopefully illustrated a bit
how this concept and implementation works.

Use-cases

Now that we had a look at how to enable this logic for a unit and how
it is implemented, let’s discuss where this actually could be useful
in real life.

  • One major benefit of dynamic user IDs is that running a
    privilege-separated service leaves no artifacts in the system. A
    system user is allocated and made use of, but it is discarded
    automatically in a safe and secure way after use, in a fashion that is
    safe for later recycling. Thus, quickly invoking a short-lived service
    for processing some job can be protected properly through a user ID
    without having to pre-allocate it and without this draining the
    available UID pool any longer than necessary.

  • In many cases, starting a service no longer requires
    package-specific preparation. Or in other words, quite often
    useradd/mkdir/chown/chmod invocations in “post-inst” package
    scripts, as well as
    sysusers.d
    and
    tmpfiles.d
    drop-ins become unnecessary, as the DynamicUser= and
    StateDirectory=/CacheDirectory=/LogsDirectory= logic can do the
    necessary work automatically, on-demand and with a well-defined
    life-cycle.

  • By combining dynamic user IDs with the transient unit concept, new
    creative ways of sand-boxing are made available. For example, let’s say
    you don’t trust the correct implementation of the sort command. You
    can now lock it into a simple, robust, dynamic UID sandbox with a
    simple systemd-run and still integrate it into a shell pipeline like
    any other command. Here’s an example, showcasing a shell pipeline
    whose middle element runs as a dynamically on-the-fly allocated UID,
    that is released when the pipelines ends.

    # cat some-file.txt | systemd-run ---pipe --property=DynamicUser=1 sort -u | grep -i foobar > some-other-file.txt
    
  • By combining dynamic user IDs with the systemd templating logic it
    is now possible to do much more fine-grained and fully automatic UID
    management. For example, let’s say you have a template unit file
    /etc/systemd/system/[email protected]:

    [Service]
    ExecStart=/usr/bin/myfoobarserviced
    DynamicUser=1
    StateDirectory=foobar/%i
    

    Now, let’s say you want to start one instance of this service for
    each of your customers. All you need to do now for that is:

    # systemctl enable [email protected] --now
    

    And you are done. (Invoke this as many times as you like, each time
    replacing customerxyz by some customer identifier, you get the
    idea.)

  • By combining dynamic user IDs with socket activation you may easily
    implement a system where each incoming connection is served by a
    process instance running as a different, fresh, newly allocated UID
    within its own sandbox. Here’s an example waldo.socket:

    [Socket]
    ListenStream=2048
    Accept=yes
    

    With a matching [email protected]:

    [Service]
    ExecStart=-/usr/bin/myservicebinary
    DynamicUser=yes
    

    With the two unit files above, systemd will listen on TCP/IP port
    2048, and for each incoming connection invoke a fresh instance of
    [email protected], each time utilizing a different, new,
    dynamically allocated UID, neatly isolated from any other
    instance.

  • Dynamic user IDs combine very well with state-less systems,
    i.e. systems that come up with an unpopulated /etc and /var. A
    service using dynamic user IDs and the StateDirectory=,
    CacheDirectory=, LogsDirectory= and RuntimeDirectory= concepts
    will implicitly allocate the users and directories it needs for
    running, right at the moment where it needs it.

Dynamic users are a very generic concept, hence a multitude of other
uses are thinkable; the list above is just supposed to trigger your
imagination.

What does this mean for you as a packager?

I am pretty sure that a large number of services shipped with today’s
distributions could benefit from using DynamicUser= and
StateDirectory= (and related settings). It often allows removal of
post-inst packaging scripts altogether, as well as any sysusers.d
and tmpfiles.d drop-ins by unifying the needed declarations in the
unit file itself. Hence, as a packager please consider switching your
unit files over. That said, there are a number of conditions where
DynamicUser= and StateDirectory= (and friends) cannot or should
not be used. To name a few:

  1. Service that need to write to files outside of /run/<package>,
    /var/lib/<package>, /var/cache/<package>, /var/log/<package>,
    /var/tmp, /tmp, /dev/shm are generally incompatible with this
    scheme. This rules out daemons that upgrade the system as one example,
    as that involves writing to /usr.

  2. Services that maintain a herd of processes with different user
    IDs. Some SMTP services are like this. If your service has such a
    super-server design, UID management needs to be done by the
    super-server itself, which rules out systemd doing its dynamic UID
    magic for it.

  3. Services which run as root (obviously…) or are otherwise
    privileged.

  4. Services that need to live in the same mount name-space as the host
    system (for example, because they want to establish mount points
    visible system-wide). As mentioned DynamicUser= implies
    ProtectSystem=, PrivateTmp= and related options, which all require
    the service to run in its own mount name-space.

  5. Your focus is older distributions, i.e. distributions that do not
    have systemd 232 (for DynamicUser=) or systemd 235 (for
    StateDirectory= and friends) yet.

  6. If your distribution’s packaging guides don’t allow it. Consult
    your packaging guides, and possibly start a discussion on your
    distribution’s mailing list about this.

Notes

A couple of additional, random notes about the implementation and use
of these features:

  1. Do note that allocating or deallocating a dynamic user leaves
    /etc/passwd untouched. A dynamic user is added into the user
    database through the glibc NSS module
    nss-systemd,
    and this information never hits the disk.

  2. On traditional UNIX systems it was the job of the daemon process
    itself to drop privileges, while the DynamicUser= concept is
    designed around the service manager (i.e. systemd) being responsible
    for that. That said, since v235 there’s a way to marry DynamicUser=
    and such services which want to drop privileges on their own. For
    that, turn on DynamicUser= and set
    User=
    to the user name the service wants to setuid() to. This has the
    effect that systemd will allocate the dynamic user under the specified
    name when the service is started. Then, prefix the command line you
    specify in
    ExecStart=
    with a single ! character. If you do, the user is allocated for the
    service, but the daemon binary is is invoked as root instead of the
    allocated user, under the assumption that the daemon changes its UID
    on its own the right way. Not that after registration the user will
    show up instantly in the user database, and is hence resolvable like
    any other by the daemon process. Example:
    ExecStart=!/usr/bin/mydaemond

  3. You may wonder why systemd uses the UID range 61184–65519 for its
    dynamic user allocations (side note: in hexadecimal this reads as
    0xEF00–0xFFEF). That’s because distributions (specifically Fedora)
    tend to allocate regular users from below the 60000 range, and we
    don’t want to step into that. We also want to stay away from 65535 and
    a bit around it, as some of these UIDs have special meanings (65535 is
    often used as special value for “invalid” or “no” UID, as it is
    identical to the 16bit value -1; 65534 is generally mapped to the
    “nobody” user, and is where some kernel subsystems map unmappable
    UIDs). Finally, we want to stay within the 16bit range. In a user
    name-spacing world each container tends to have much less than the full
    32bit UID range available that Linux kernels theoretically
    provide. Everybody apparently can agree that a container should at
    least cover the 16bit range though — already to include a nobody
    user. (And quite frankly, I am pretty sure assigning 64K UIDs per
    container is nicely systematic, as the the higher 16bit of the 32bit
    UID values this way become a container ID, while the lower 16bit
    become the logical UID within each container, if you still follow what
    I am babbling here…). And before you ask: no this range cannot be
    changed right now, it’s compiled in. We might change that eventually
    however.

  4. You might wonder what happens if you already used UIDs from the
    61184–65519 range on your system for other purposes. systemd should
    handle that mostly fine, as long as that usage is properly registered
    in the user database: when allocating a dynamic user we pick a UID,
    see if it is currently used somehow, and if yes pick a different one,
    until we find a free one. Whether a UID is used right now or not is
    checked through NSS calls. Moreover the IPC object lists are checked to
    see if there are any objects owned by the UID we are about to
    pick. This means systemd will avoid using UIDs you have assigned
    otherwise. Note however that this of course makes the pool of
    available UIDs smaller, and in the worst cases this means that
    allocating a dynamic user might fail because there simply are no
    unused UIDs in the range.

  5. If not specified otherwise the name for a dynamically allocated
    user is derived from the service name. Not everything that’s valid in
    a service name is valid in a user-name however, and in some cases a
    randomized name is used instead to deal with this. Often it makes
    sense to pick the user names to register explicitly. For that use
    User= and choose whatever you like.

  6. If you pick a user name with User= and combine it with
    DynamicUser= and the user already exists statically it will be used
    for the service and the dynamic user logic is automatically
    disabled. This permits automatic up- and downgrades between static and
    dynamic UIDs. For example, it provides a nice way to move a system
    from static to dynamic UIDs in a compatible way: as long as you select
    the same User= value before and after switching DynamicUser= on,
    the service will continue to use the statically allocated user if it
    exists, and only operates in the dynamic mode if it does not. This is
    useful for other cases as well, for example to adapt a service that
    normally would use a dynamic user to concepts that require statically
    assigned UIDs, for example to marry classic UID-based file system
    quota with such services.

  7. systemd always allocates a pair of dynamic UID and GID at the same
    time, with the same numeric ID.

  8. If the Linux kernel had a “shiftfs” or similar functionality,
    i.e. a way to mount an existing directory to a second place, but map
    the exposed UIDs/GIDs in some way configurable at mount time, this
    would be excellent for the implementation of StateDirectory= in
    conjunction with DynamicUser=. It would make the recursive
    chown()ing step unnecessary, as the host version of the state
    directory could simply be mounted into a the service’s mount
    name-space, with a shift applied that maps the directory’s owner to the
    services’ UID/GID. But I don’t have high hopes in this regard, as all
    work being done in this area appears to be bound to user name-spacing
    — which is a concept not used here (and I guess one could say user
    name-spacing is probably more a source of problems than a solution to
    one, but you are welcome to disagree on that).

And that’s all for now. Enjoy your dynamic users!

Things Go Better With Step Functions

Post Syndicated from Jeff Barr original https://aws.amazon.com/blogs/aws/things-go-better-with-step-functions/

I often give presentations on Amazon’s culture of innovation, and start out with a slide that features a revealing quote from Amazon founder Jeff Bezos:

I love to sit down with our customers and to learn how we have empowered their creativity and to pursue their dreams. Earlier this year I chatted with Patrick from The Coca-Cola Company in order to learn how they used AWS Step Functions and other AWS services to support the Coke.com Vending Pass program. This program includes drink rewards earned by purchasing products at vending machines equipped to support mobile payments using the Coca-Cola Vending Pass. Participants swipe their NFC-enabled phones to complete an Apple Pay or Android Pay purchase, identifying themselves to the vending machine and earning credit towards future free vending purchases in the process

After the swipe, a combination of SNS topics and AWS Lambda functions initiated a pair of calls to some existing backend code to count the vending points and update the participant’s record. Unfortunately, the backend code was slow to react and had some timing dependencies, leading to missing updates that had the potential to confuse Vending Pass participants. The initial solution to this issue was very simple: modify the Lambda code to include a 90 second delay between the two calls. This solved the problem, but ate up process time for no good reason (billing for the use of Lambda functions is based on the duration of the request, in 100 ms intervals).

In order to make their solution more cost-effective, the team turned to AWS Step Functions, building a very simple state machine. As I wrote in an earlier blog post, Step Functions coordinate the components of distributed applications and microservices at scale, using visual workflows that are easy to build.

Coke built a very simple state machine to simplify their business logic and reduce their costs. Yours can be equally simple, or they can make use of other Step Function features such as sequential and parallel execution and the ability to make decisions and choose alternate states. The Coke state machine looks like this:

The FirstState and the SecondState states (Task states) call the appropriate Lambda functions while Step Functions implements the 90 second delay (a Wait state). This modification simplified their logic and reduced their costs. Here’s how it all fits together:

 

What’s Next
This initial success led them to take a closer look at serverless computing and to consider using it for other projects. Patrick told me that they have already seen a boost in productivity and developer happiness. Developers no longer need to wait for servers to be provisioned, and can now (as Jeff says) unleash their creativity and pursue their dreams. They expect to use Step Functions to improve the scalability, functionality, and reliability of their applications, going far beyond the initial use for the Coca-Cola Vending Pass. For example, Coke has built a serverless solution for publishing nutrition information to their food service partners using Lambda, Step Functions, and API Gateway.

Patrick and his team are now experimenting with machine learning and artificial intelligence. They built a prototype application to analyze a stream of photos from Instagram and extract trends in tastes and flavors. The application (built as a quick, one-day prototype) made use of Lambda, Amazon DynamoDB, Amazon API Gateway, and Amazon Rekognition and was, in Patrick’s words, a “big win and an enabler.”

In order to build serverless applications even more quickly, the development team has created an internal CI/CD reference architecture that builds on the Serverless Application Framework. The architecture includes a guided tour of Serverless and some boilerplate code to access internal services and assets. Patrick told me that this model allows them to easily scale promising projects from “a guy with a computer” to an entire development team.

Patrick will be on stage at AWS re:Invent next to my colleague Tim Bray. To meet them in person, be sure to attend SRV306 – State Machines in the Wild! How Customers Use AWS Step Functions.

Jeff;

Natural Language Processing at Clemson University – 1.1 Million vCPUs & EC2 Spot Instances

Post Syndicated from Jeff Barr original https://aws.amazon.com/blogs/aws/natural-language-processing-at-clemson-university-1-1-million-vcpus-ec2-spot-instances/

My colleague Sanjay Padhi shared the guest post below in order to recognize an important milestone in the use of EC2 Spot Instances.

Jeff;


A group of researchers from Clemson University achieved a remarkable milestone while studying topic modeling, an important component of machine learning associated with natural language processing, breaking the record for creating the largest high-performance cluster by using more than 1,100,000 vCPUs on Amazon EC2 Spot Instances running in a single AWS region. The researchers conducted nearly half a million topic modeling experiments to study how human language is processed by computers. Topic modeling helps in discovering the underlying themes that are present across a collection of documents. Topic models are important because they are used to forecast business trends and help in making policy or funding decisions. These topic models can be run with many different parameters and the goal of the experiments is to explore how these parameters affect the model outputs.

The Experiment
Professor Amy Apon, Co-Director of the Complex Systems, Analytics and Visualization Institute at Clemson University with Professor Alexander Herzog and graduate students Brandon Posey and Christopher Gropp in collaboration with members of the AWS team as well as AWS Partner Omnibond performed the experiments.  They used software infrastructure based on CloudyCluster that provisions high performance computing clusters on dynamically allocated AWS resources using Amazon EC2 Spot Fleet. Spot Fleet is a collection of biddable spot instances in EC2 responsible for maintaining a target capacity specified during the request. The SLURM scheduler was used as an overlay virtual workload manager for the data analytics workflows. The team developed additional provisioning and workflow automation software as shown below for the design and orchestration of the experiments. This setup allowed them to evaluate various topic models on different data sets with massively parallel parameter sweeps on dynamically allocated AWS resources. This framework can easily be used beyond the current study for other scientific applications that use parallel computing.

Ramping to 1.1 Million vCPUs
The figure below shows elastic, automatic expansion of resources as a function of time, in the US East (Northern Virginia) Region. At just after 21:40 (GMT-1) on Aug. 26, 2017, the number of vCPUs utilized was 1,119,196. Clemson researchers also took advantage of the new per-second billing for the EC2 instances that they launched. The vCPU count usage is comparable to the core count on the largest supercomputers in the world.

Here’s the breakdown of the EC2 instance types that they used:

Campus resources at Clemson funded by the National Science Foundation were used to determine an effective configuration for the AWS experiments as compared to campus resources, and the AWS cloud resources complement the campus resources for large-scale experiments.

Meet the Team
Here’s the team that ran the experiment (Professor Alexander Herzog, graduate students Christopher Gropp and Brandon Posey, and Professor Amy Apon):

Professor Apon said about the experiment:

I am absolutely thrilled with the outcome of this experiment. The graduate students on the project are amazing. They used resources from AWS and Omnibond and developed a new software infrastructure to perform research at a scale and time-to-completion not possible with only campus resources. Per-second billing was a key enabler of these experiments.

Boyd Wilson (CEO, Omnibond, member of the AWS Partner Network) told me:

Participating in this project was exciting, seeing how the Clemson team developed a provisioning and workflow automation tool that tied into CloudyCluster to build a huge Spot Fleet supercomputer in a single region in AWS was outstanding.

About the Experiment
The experiments test parameter combinations on a range of topics and other parameters used in the topic model. The topic model outputs are stored in Amazon S3 and are currently being analyzed. The models have been applied to 17 years of computer science journal abstracts (533,560 documents and 32,551,540 words) and full text papers from the NIPS (Neural Information Processing Systems) Conference (2,484 documents and 3,280,697 words). This study allows the research team to systematically measure and analyze the impact of parameters and model selection on model convergence, topic composition and quality.

Looking Forward
This study constitutes an interaction between computer science, artificial intelligence, and high performance computing. Papers describing the full study are being submitted for peer-reviewed publication. I hope that you enjoyed this brief insight into the ways in which AWS is helping to break the boundaries in the frontiers of natural language processing!

Sanjay Padhi, Ph.D, AWS Research and Technical Computing

 

Using AWS CodePipeline, AWS CodeBuild, and AWS Lambda for Serverless Automated UI Testing

Post Syndicated from Prakash Palanisamy original https://aws.amazon.com/blogs/devops/using-aws-codepipeline-aws-codebuild-and-aws-lambda-for-serverless-automated-ui-testing/

Testing the user interface of a web application is an important part of the development lifecycle. In this post, I’ll explain how to automate UI testing using serverless technologies, including AWS CodePipeline, AWS CodeBuild, and AWS Lambda.

I built a website for UI testing that is hosted in S3. I used Selenium to perform cross-browser UI testing on Chrome, Firefox, and PhantomJS, a headless WebKit browser with Ghost Driver, an implementation of the WebDriver Wire Protocol. I used Python to create test cases for ChromeDriver, FirefoxDriver, or PhatomJSDriver based the browser against which the test is being executed.

Resources referred to in this post, including the AWS CloudFormation template, test and status websites hosted in S3, AWS CodeBuild build specification files, AWS Lambda function, and the Python script that performs the test are available in the serverless-automated-ui-testing GitHub repository.

S3 Hosted Test Website:

AWS CodeBuild supports custom containers so we can use the Selenium/standalone-Firefox and Selenium/standalone-Chrome containers, which include prebuild Firefox and Chrome browsers, respectively. Xvfb performs the graphical operation in virtual memory without any display hardware. It will be installed in the CodeBuild containers during the install phase.

Build Spec for Chrome and Firefox

The build specification for Chrome and Firefox testing includes multiple phases:

  • The environment variables section contains a set of default variables that are overridden while creating the build project or triggering the build.
  • As part of install phase, required packages like Xvfb and Selenium are installed using yum.
  • During the pre_build phase, the test bed is prepared for test execution.
  • During the build phase, the appropriate DISPLAY is set and the tests are executed.
version: 0.2

env:
  variables:
    BROWSER: "chrome"
    WebURL: "https://sampletestweb.s3-eu-west-1.amazonaws.com/website/index.html"
    ArtifactBucket: "codebuild-demo-artifact-repository"
    MODULES: "mod1"
    ModuleTable: "test-modules"
    StatusTable: "blog-test-status"

phases:
  install:
    commands:
      - apt-get update
      - apt-get -y upgrade
      - apt-get install xvfb python python-pip build-essential -y
      - pip install --upgrade pip
      - pip install selenium
      - pip install awscli
      - pip install requests
      - pip install boto3
      - cp xvfb.init /etc/init.d/xvfb
      - chmod +x /etc/init.d/xvfb
      - update-rc.d xvfb defaults
      - service xvfb start
      - export PATH="$PATH:`pwd`/webdrivers"
  pre_build:
    commands:
      - python prepare_test.py
  build:
    commands:
      - export DISPLAY=:5
      - cd tests
      - echo "Executing simple test..."
      - python testsuite.py

Because Ghost Driver runs headless, it can be executed on AWS Lambda. In keeping with a fire-and-forget model, I used CodeBuild to create the PhantomJS Lambda function and trigger the test invocations on Lambda in parallel. This is powerful because many tests can be executed in parallel on Lambda.

Build Spec for PhantomJS

The build specification for PhantomJS testing also includes multiple phases. It is a little different from the preceding example because we are using AWS Lambda for the test execution.

  • The environment variables section contains a set of default variables that are overridden while creating the build project or triggering the build.
  • As part of install phase, the required packages like Selenium and the AWS CLI are installed using yum.
  • During the pre_build phase, the test bed is prepared for test execution.
  • During the build phase, a zip file that will be used to create the PhantomJS Lambda function is created and tests are executed on the Lambda function.
version: 0.2

env:
  variables:
    BROWSER: "phantomjs"
    WebURL: "https://sampletestweb.s3-eu-west-1.amazonaws.com/website/index.html"
    ArtifactBucket: "codebuild-demo-artifact-repository"
    MODULES: "mod1"
    ModuleTable: "test-modules"
    StatusTable: "blog-test-status"
    LambdaRole: "arn:aws:iam::account-id:role/role-name"

phases:
  install:
    commands:
      - apt-get update
      - apt-get -y upgrade
      - apt-get install python python-pip build-essential -y
      - apt-get install zip unzip -y
      - pip install --upgrade pip
      - pip install selenium
      - pip install awscli
      - pip install requests
      - pip install boto3
  pre_build:
    commands:
      - python prepare_test.py
  build:
    commands:
      - cd lambda_function
      - echo "Packaging Lambda Function..."
      - zip -r /tmp/lambda_function.zip ./*
      - func_name=`echo $CODEBUILD_BUILD_ID | awk -F ':' '{print $1}'`-phantomjs
      - echo "Creating Lambda Function..."
      - chmod 777 phantomjs
      - |
         func_list=`aws lambda list-functions | grep FunctionName | awk -F':' '{print $2}' | tr -d ', "'`
         if echo "$func_list" | grep -qw $func_name
         then
             echo "Lambda function already exists."
         else
             aws lambda create-function --function-name $func_name --runtime "python2.7" --role $LambdaRole --handler "testsuite.lambda_handler" --zip-file fileb:///tmp/lambda_function.zip --timeout 150 --memory-size 1024 --environment Variables="{WebURL=$WebURL, StatusTable=$StatusTable}" --tags Name=$func_name
         fi
      - export PhantomJSFunction=$func_name
      - cd ../tests/
      - python testsuite.py

The list of test cases and the test modules that belong to each case are stored in an Amazon DynamoDB table. Based on the list of modules passed as an argument to the CodeBuild project, CodeBuild gets the test cases from that table and executes them. The test execution status and results are stored in another Amazon DynamoDB table. It will read the test status from the status table in DynamoDB and display it.

AWS CodeBuild and AWS Lambda perform the test execution as individual tasks. AWS CodePipeline plays an important role here by enabling continuous delivery and parallel execution of tests for optimized testing.

Here’s how to do it:

In AWS CodePipeline, create a pipeline with four stages:

  • Source (AWS CodeCommit)
  • UI testing (AWS Lambda and AWS CodeBuild)
  • Approval (manual approval)
  • Production (AWS Lambda)

Pipeline stages, the actions in each stage, and transitions between stages are shown in the following diagram.

This design implemented in AWS CodePipeline looks like this:

CodePipeline automatically detects a change in the source repository and triggers the execution of the pipeline.

In the UITest stage, there are two parallel actions:

  • DeployTestWebsite invokes a Lambda function to deploy the test website in S3 as an S3 website.
  • DeployStatusPage invokes another Lambda function to deploy in parallel the status website in S3 as an S3 website.

Next, there are three parallel actions that trigger the CodeBuild project:

  • TestOnChrome launches a container to perform the Selenium tests on Chrome.
  • TestOnFirefox launches another container to perform the Selenium tests on Firefox.
  • TestOnPhantomJS creates a Lambda function and invokes individual Lambda functions per test case to execute the test cases in parallel.

You can monitor the status of the test execution on the status website, as shown here:

When the UI testing is completed successfully, the pipeline continues to an Approval stage in which a notification is sent to the configured SNS topic. The designated team member reviews the test status and approves or rejects the deployment. Upon approval, the pipeline continues to the Production stage, where it invokes a Lambda function and deploys the website to a production S3 bucket.

I used a CloudFormation template to set up my continuous delivery pipeline. The automated-ui-testing.yaml template, available from GitHub, sets up a full-featured pipeline.

When I use the template to create my pipeline, I specify the following:

  • AWS CodeCommit repository.
  • SNS topic to send approval notification.
  • S3 bucket name where the artifacts will be stored.

The stack name should follow the rules for S3 bucket naming because it will be part of the S3 bucket name.

When the stack is created successfully, the URLs for the test website and status website appear in the Outputs section, as shown here:

Conclusion

In this post, I showed how you can use AWS CodePipeline, AWS CodeBuild, AWS Lambda, and a manual approval process to create a continuous delivery pipeline for serverless automated UI testing. Websites running on Amazon EC2 instances or AWS Elastic Beanstalk can also be tested using similar approach.


About the author

Prakash Palanisamy is a Solutions Architect for Amazon Web Services. When he is not working on Serverless, DevOps or Alexa, he will be solving problems in Project Euler. He also enjoys watching educational documentaries.

Prime Day 2017 – Powered by AWS

Post Syndicated from Jeff Barr original https://aws.amazon.com/blogs/aws/prime-day-2017-powered-by-aws/

The third annual Prime Day set another round of records for global orders, topping Black Friday and Cyber Monday, making it the biggest day in Amazon retail history. Over the course of the 30 hour event, tens of millions of Prime members purchased things like Echo Dots, Fire tablets, programmable pressure cookers, espresso machines, rechargeable batteries, and much more! July 11th also set a record for the number of new Prime memberships, as people signed up in order to take advantage of hundreds of thousands of deals. Amazon customers shopped online and made heavy use of the Amazon App, with mobile orders more than doubling from last Prime Day.

Powered by AWS
Last year I told you about How AWS Powered Amazon’s Biggest Day Ever, and shared what the team had learned with regard to preparation, automation, monitoring, and thinking big. All of those lessons still apply and you can read that post to learn more. Preparation for this year’s Prime Day (which started just days after Prime Day 2016 wrapped up) started by collecting and sharing best practices and identifying areas for improvement, proceeding to implementation and stress testing as the big day approached. Two of the best practices involve auditing and GameDay:

Auditing – This is a formal way for us to track preparations, identify risks, and to track progress against our objectives. Each team must respond to a series of detailed technical and operational questions that are designed to help them determine their readiness. On the technical side, questions could revolve around time to recovery after a database failure, including the all-important check of the TTL (time to live) for the CNAME. Operational questions address schedules for on-call personnel, points of contact, and ownership of services & instances.

GameDay – This practice (which I believe originated with former Amazonian Jesse Robbins), is intended to validate all of the capacity planning & preparation and to verify that all of the necessary operational practices are in place and work as expected. It introduces simulated failures and helps to train the team to identify and quickly resolve issues, building muscle memory in the process. It also tests failover and recovery capabilities, and can expose latent defects that are lurking under the covers. GameDays help teams to understand scaling drivers (page views, orders, and so forth) and gives them an opportunity to test their scaling practices. To learn more, read Resilience Engineering: Learning to Embrace Failure or watch the video: GameDay: Creating Resiliency Through Destruction.

Prime Day 2017 Metrics
So, how did we do this year?

The AWS teams checked their dashboards and log files, and were happy to share their metrics with me. Here are a few of the most interesting ones:

Block Storage – Use of Amazon Elastic Block Store (EBS) grew by 40% year-over-year, with aggregate data transfer jumping to 52 petabytes (a 50% increase) for the day and total I/O requests rising to 835 million (a 30% increase). The team told me that they loved the elasticity of EBS, and that they were able to ramp down on capacity after Prime Day concluded instead of being stuck with it.

NoSQL Database – Amazon DynamoDB requests from Alexa, the Amazon.com sites, and the Amazon fulfillment centers totaled 3.34 trillion, peaking at 12.9 million per second. According to the team, the extreme scale, consistent performance, and high availability of DynamoDB let them meet needs of Prime Day without breaking a sweat.

Stack Creation – Nearly 31,000 AWS CloudFormation stacks were created for Prime Day in order to bring additional AWS resources on line.

API Usage – AWS CloudTrail processed over 50 billion events and tracked more than 419 billion calls to various AWS APIs, all in support of Prime Day.

Configuration TrackingAWS Config generated over 14 million Configuration items for AWS resources.

You Can Do It
Running an event that is as large, complex, and mission-critical as Prime Day takes a lot of planning. If you have an event of this type in mind, please take a look at our new Infrastructure Event Readiness white paper. Inside, you will learn how to design and provision your applications to smoothly handle planned scaling events such as product launches or seasonal traffic spikes, with sections on automation, resiliency, cost optimization, event management, and more.

Jeff;

 

Disabling Intel Hyper-Threading Technology on Amazon EC2 Windows Instances

Post Syndicated from Brian Beach original https://aws.amazon.com/blogs/compute/disabling-intel-hyper-threading-technology-on-amazon-ec2-windows-instances/

In a prior post, Disabling Intel Hyper-Threading on Amazon Linux, I investigated how the Linux kernel enumerates CPUs. I also discussed the options to disable Intel Hyper-Threading (HT Technology) in Amazon Linux running on Amazon EC2.

In this post, I do the same for Microsoft Windows Server 2016 running on EC2 instances. I begin with a quick review of HT Technology and the reasons you might want to disable it. I also recommend that you take a moment to review the prior post for a more thorough foundation.

HT Technology

HT Technology makes a single physical processor appear as multiple logical processors. Each core in an Intel Xeon processor has two threads of execution. Most of the time, these threads can progress independently; one thread executing while the other is waiting on a relatively slow operation (for example, reading from memory) to occur. However, the two threads do share resources and occasionally one thread is forced to wait while the other is executing.

There a few unique situations where disabling HT Technology can improve performance. One example is high performance computing (HPC) workloads that rely heavily on floating point operations. In these rare cases, it can be advantageous to disable HT Technology. However, these cases are rare, and for the overwhelming majority of workloads you should leave it enabled. I recommend that you test with and without HT Technology enabled, and only disable threads if you are sure it will improve performance.

Exploring HT Technology on Microsoft Windows

Here’s how Microsoft Windows enumerates CPUs. As before, I am running these examples on an m4.2xlarge. I also chose to run Windows Server 2016, but you can walk through these exercises on any version of Windows. Remember that the m4.2xlarge has eight vCPUs, and each vCPU is a thread of an Intel Xeon core. Therefore, the m4.2xlarge has four cores, each of which run two threads, resulting in eight vCPUs.

Windows does not have a built-in utility to examine CPU configuration, but you can download the Sysinternals coreinfo utility from Microsoft’s website. This utility provides useful information about the system CPU and memory topology. For this walkthrough, you enumerate the individual CPUs, which you can do by running coreinfo -c. For example:

C:\Users\Administrator >coreinfo -c

Coreinfo v3.31 - Dump information on system CPU and memory topology
Copyright (C) 2008-2014 Mark Russinovich
Sysinternals - www.sysinternals.com

Logical to Physical Processor Map:
**------ Physical Processor 0 (Hyperthreaded)
--**---- Physical Processor 1 (Hyperthreaded)
----**-- Physical Processor 2 (Hyperthreaded)
------** Physical Processor 3 (Hyperthreaded)

As you can see from the screenshot, the coreinfo utility displays a table where each row is a physical core and each column is a logical CPU. In other words, the two asterisks on the first line indicate that CPU 0 and CPU 1 are the two threads in the first physical core. Therefore, my m4.2xlarge has for four physical processors and each processor has two threads resulting in eight total CPUs, just as expected.

It is interesting to note that Windows Server 2016 enumerates CPUs in a different order than Linux. Remember from the prior post that Linux enumerated the first thread in each core, followed by the second thread in each core. You can see from the output earlier that Windows Server 2016, enumerates both threads in the first core, then both threads in the second core, and so on. The diagram below shows the relationship of CPUs to cores and threads in both operating systems.

In the Linux post, I disabled CPUs 4–6, leaving one thread per core, and effectively disabling HT Technology. You can see from the diagram that you must disable the odd-numbered threads (that is, 1, 3, 5, and 7) to achieve the same result in Windows. Here’s how to do that.

Disabling HT Technology on Microsoft Windows

In Linux, you can globally disable CPUs dynamically. In Windows, there is no direct equivalent that I could find, but there are a few alternatives.

First, you can disable CPUs using the msconfig.exe tool. If you choose Boot, Advanced Options, you have the option to set the number of processors. In the example below, I limit my m4.2xlarge to four CPUs. Restart for this change to take effect.

Unfortunately, Windows does not disable hyperthreaded CPUs first and then real cores, as Linux does. As you can see in the following output, coreinfo reports that my c4.2xlarge has two real cores and four hyperthreads, after rebooting. Msconfig.exe is useful for disabling cores, but it does not allow you to disable HT Technology.

Note: If you have been following along, you can re-enable all your CPUs by unselecting the Number of processors check box and rebooting your system.

 

C:\Users\Administrator >coreinfo -c

Coreinfo v3.31 - Dump information on system CPU and memory topology
Copyright (C) 2008-2014 Mark Russinovich
Sysinternals - www.sysinternals.com

Logical to Physical Processor Map:
**-- Physical Processor 0 (Hyperthreaded)
--** Physical Processor 1 (Hyperthreaded)

While you cannot disable HT Technology systemwide, Windows does allow you to associate a particular process with one or more CPUs. Microsoft calls this, “processor affinity”. To see an example, use the following steps.

  1. Launch an instance of Notepad.
  2. Open Windows Task Manager and choose Processes.
  3. Open the context (right click) menu on notepad.exe and choose Set Affinity….

This brings up the Processor Affinity dialog box.

As you can see, all the CPUs are allowed to run this instance of notepad.exe. You can uncheck a few CPUs to exclude them. Windows is smart enough to allow any scheduled operations to continue to completion on disabled CPUs. It then saves its state at the next scheduling event, and resumes those operations on another CPU. To ensure that only one thread in each core is able to run a process, you uncheck every other core. This effectively disables HT Technology for this process. For example:

Of course, this can be tedious when you have a large number of cores. Remember that the x1.32xlarge has 128 CPUs. Luckily, you can set the affinity of a running process from PowerShell using the Get-Process cmdlet. For example:

PS C:\&gt; (Get-Process -Name 'notepad').ProcessorAffinity = 0x55;

The ProcessorAffinity attribute takes a bitmask in hexadecimal format. 0x55 in hex is equivalent to 01010101 in binary. Think of the binary encoding as 1=enabled and 0=disabled. This is slightly confusing, but we work left to right so that CPU 0 is the rightmost bit and CPU 7 is the leftmost bit. Therefore, 01010101 means that the first thread in each CPU is enabled just as it was in the diagram earlier.

The calculator built into Windows includes a “programmer view” that helps you convert from hexadecimal to binary. In addition, the ProcessorAffinity attribute is a 64-bit number. Therefore, you can only configure the processor affinity on systems up to 64 CPUs. At the moment, only the x1.32xlarge has more than 64 vCPUs.

In the preceding examples, you changed the processor affinity of a running process. Sometimes, you want to start a process with the affinity already configured. You can do this using the start command. The start command includes an affinity flag that takes a hexadecimal number like the PowerShell example earlier.

C:\Users\Administrator&gt;start /affinity 55 notepad.exe

It is interesting to note that a child process inherits the affinity from its parent. For example, the following commands create a batch file that launches Notepad, and starts the batch file with the affinity set. If you examine the instance of Notepad launched by the batch file, you see that the affinity has been applied to as well.

C:\Users\Administrator&gt;echo notepad.exe > test.bat
C:\Users\Administrator&gt;start /affinity 55 test.bat

This means that you can set the affinity of your task scheduler and any tasks that the scheduler starts inherits the affinity. So, you can disable every other thread when you launch the scheduler and effectively disable HT Technology for all of the tasks as well. Be sure to test this point, however, as some schedulers override the normal inheritance behavior and explicitly set processor affinity when starting a child process.

Conclusion

While the Windows operating system does not allow you to disable logical CPUs, you can set processor affinity on individual processes. You also learned that Windows Server 2016 enumerates CPUs in a different order than Linux. Therefore, you can effectively disable HT Technology by restricting a process to every other CPU. Finally, you learned how to set affinity of both new and running processes using Task Manager, PowerShell, and the start command.

Note: this technical approach has nothing to do with control over software licensing, or licensing rights, which are sometimes linked to the number of “CPUs” or “cores.” For licensing purposes, those are legal terms, not technical terms. This post did not cover anything about software licensing or licensing rights.

If you have questions or suggestions, please comment below.

Implement Continuous Integration and Delivery of Apache Spark Applications using AWS

Post Syndicated from Luis Caro Perez original https://aws.amazon.com/blogs/big-data/implement-continuous-integration-and-delivery-of-apache-spark-applications-using-aws/

When you develop Apache Spark–based applications, you might face some additional challenges when dealing with continuous integration and deployment pipelines, such as the following common issues:

  • Applications must be tested on real clusters using automation tools (live test)
  • Any user or developer must be able to easily deploy and use different versions of both the application and infrastructure to be able to debug, experiment on, and test different functionality.
  • Infrastructure needs to be evaluated and tested along with the application that uses it.

In this post, we walk you through a solution that implements a continuous integration and deployment pipeline supported by AWS services. The pipeline offers the following workflow:

  • Deploy the application to a QA stage after a commit is performed to the source code.
  • Perform a unit test using Spark local mode.
  • Deploy to a dynamically provisioned Amazon EMR cluster and test the Spark application on it
  • Update the application as an AWS Service Catalog product version, allowing a user to deploy any version (commit) of the application on demand.

Solution overview

The following diagram shows the pipeline workflow.

The solution uses AWS CodePipeline, which allows users to orchestrate and automate the build, test, and deploy stages for application source code. The solution consists of a pipeline that contains the following stages:

  • Source: Both the Spark application source code in addition to the AWS CloudFormation template file for deploying the application are committed to version control. In this example, we use AWS CodeCommit. For an example of the application source code, see zip. 
  • Build: In this stage, you use Apache Maven both to generate the application .jar binaries and to execute all of the application unit tests that end with *Spec.scala. In this example, we use AWS CodeBuild, which runs the unit tests given that they are designed to use Spark local mode.
  • QADeploy: In this stage, the .jar file built previously is deployed using the CloudFormation template included with the application source code. All the resources are created in this stage, such as networks, EMR clusters, and so on. 
  • LiveTest: In this stage, you use Apache Maven to execute all the application tests that end with *SpecLive.scala. The tests submit EMR steps to the cluster created as part of the QADeploy step. The tests verify that the steps ran successfully and their results. 
  • LiveTestApproval: This stage is included in case a pipeline administrator approval is required to deploy the application to the next stages. The pipeline pauses in this stage until an administrator manually approves the release. 
  • QACleanup: In this stage, you use an AWS Lambda function to delete the CloudFormation template deployed as part of the QADeploy stage. The function does not affect any resources other than those deployed as part of the QADeploy stage. 
  • DeployProduct: In this stage, you use a Lambda function that creates or updates an AWS Service Catalog product and portfolio. Every time the pipeline releases a change to the application, the AWS Service Catalog product gets a new version, with the commit of the change as the version description. 

Try it out!

Use the provided sample template to get started using this solution. This template creates the pipeline described earlier with all of its stages. It performs an initial commit of the sample Spark application in order to trigger the first release change. To deploy the template, use the following AWS CLI command:

aws cloudformation create-stack  --template-url https://s3.amazonaws.com/aws-bigdata-blog/artifacts/sparkAppDemoForPipeline/emrSparkpipeline.yaml --stack-name emr-spark-pipeline --capabilities CAPABILITY_NAMED_IAM

After the template finishes creating resources, you see the pipeline name on the stack Outputs tab. After that, open the AWS CodePipeline console and select the newly created pipeline.

After a couple of minutes, AWS CodePipeline detects the initial commit applied by the CloudFormation stack and starts the first release.

You can watch how the pipeline goes through the Build, QADeploy, and LiveTest stages until it finally reaches the LiveTestApproval stage.

At this point, you can check the results of the test in the log files of the Build and LiveTest stage jobs on AWS CodeBuild. If you check the CloudFormation console, you see that a new template has been deployed as part of the QADeploy stage.

You can also visit the EMR console and view how the LiveTest stage submitted steps to the EMR cluster.

After performing the review, manually approve the revision on the LiveTestApproval stage by using the AWS CodePipeline console.

After the revision is approved, the pipeline proceeds to use a Lambda function that destroys the resources deployed on the QAdeploy stage. Finally, it creates or updates a product and portfolio in AWS Service Catalog. After the final stage of the pipeline is complete, you can check that the product is created successfully on the AWS Service Catalog console.

You can check the product versions and notice that the first version is the initial commit performed by the CloudFormation template.

You can proceed to share the created portfolio with any users in your AWS account and allow them to deploy any version of the Spark application. You can also perform a commit on the AWS CodeCommit repository. The pipeline is triggered automatically and repeats the pipeline execution to deploy a new version of the product.

To destroy all of the resources created by the stack, make sure all the deployed stacks using AWS Service Catalog or the QAdeploy stage are destroyed. Then, destroy the pipeline template using the following AWS CLI command:

 

aws cloudformation delete-stack --stack-name emr-spark-pipeline

Conclusion

You can use the sample template and Spark application shared in this post and adapt them for the specific needs of your own application. The pipeline can have as many stages as needed and it can be used to automatically deploy to AWS Service Catalog or a production environment using CloudFormation.

If you have questions or suggestions, please comment below.


Additional Reading

Learn how to implement authorization and auditing on Amazon EMR using Apache Ranger.

 


About the Authors

Luis Caro is a Big Data Consultant for AWS Professional Services. He works with our customers to provide guidance and technical assistance on big data projects, helping them improving the value of their solutions when using AWS.

 

 

Samuel Schmidt is a Big Data Consultant for AWS Professional Services. He works with our customers to provide guidance and technical assistance on big data projects, helping them improving the value of their solutions when using AWS.

 

 

 

AWS Cost Explorer Update – Better Filtering & Grouping, Report Management, RI Reports

Post Syndicated from Jeff Barr original https://aws.amazon.com/blogs/aws/aws-cost-explorer-update-better-filtering-grouping-report-management-ri-reports/

Our customers use Cost Explorer to better understand and manage their AWS spending, making heavy use of the reporting, analytics, and visualization tools that it provides. We launched Cost Explorer in 2014 with a focus on simplicity – single click signup, preconfigured default views, and a clean user interface (take a look back at The New AWS Cost Explorer to see where we started). The Cost Explorer has been very popular and we’ve received a lot of great feedback from our customers.

Last week we launched a major upgrade to Cost Explorer. We’ve redesigned the user interface to optimize many common workflows including filtering, report management, selection of date ranges, and grouping of data. We have also included some default reports to make it easier for you to explore the costs related to your use of Reserved Instances.

Looking at Cost Explorer
Since pictures are reportedly worth 1000 words, let’s take a closer look! Cost Explorer is part of the Billing Dashboard so I can start there:

Here’s the Billing Dashboard. I click on Cost Explorer to move ahead:

I can open up Cost Explorer or access one of three preconfigured views. I’ll go for the first option:

The default report shows my EC2 costs and usage (running hours) for the past 3 months:

I can use the Group By menu to break the costs down by EC2 instance type:

I have many other grouping options:

The filtering options are now easier to access and to edit. Here’s the full set:

I can explore my EC2 costs in any set of desired regions:

I can filter and then group by instance type to see how my spending breaks down:

I can click on Download CSV and then process the data locally:

I can also exclude certain instance types from the report. Here’s how I exclude my m4.xlarge, t2.micro, and t2.nano usage:

Report Management
Cost Explorer allows me to customize my existing reports and to create new reports from scratch. I can click on Save As to save my customized report with a new name:

I can see and manage all of my reports on the Saved Reports page (The padlock denotes a default report that cannot be edited and then overwritten):

When I click on New report I can start from a template:

After I click on Create Report, I set up my date range and filters as desired, and click on Save As. I created a report that displays my year-to-date usage of several AWS database services (Amazon Redshift, DynamoDB Accelerator (DAX), Amazon Relational Database Service (RDS), and AWS Database Migration Service):

All of my reports are accessible from the Reports menu so I can check on my costs with a click:

We also simplified the process of selecting a range of dates for a report, including options to select common date ranges:

Reserved Instance Reports
Cost Explorer also includes a pair of reports that will help you to understand and optimize your usage of Reserved Instances. I don’t own an RI’s so I used screen shots supplied by the team.

The RI Utilization report allows you to see how much of your purchased RI capacity is being put to use (the dashed red line represents a utilization target that you can specify):

The RI Coverage report tells you how much of your EC2 usage is being handled by Reserved Instances (this time, the dashed red line represents the desired amount of coverage):

I hope you have enjoyed this tour of the updated Cost Explorer. It is available now and you can start using it today!

Jeff;

Announcing the Winners of the AWS Chatbot Challenge – Conversational, Intelligent Chatbots using Amazon Lex and AWS Lambda

Post Syndicated from Tara Walker original https://aws.amazon.com/blogs/aws/announcing-the-winners-of-the-aws-chatbot-challenge-conversational-intelligent-chatbots-using-amazon-lex-and-aws-lambda/

A couple of months ago on the blog, I announced the AWS Chatbot Challenge in conjunction with Slack. The AWS Chatbot Challenge was an opportunity to build a unique chatbot that helped to solve a problem or that would add value for its prospective users. The mission was to build a conversational, natural language chatbot using Amazon Lex and leverage Lex’s integration with AWS Lambda to execute logic or data processing on the backend.

I know that you all have been anxiously waiting to hear announcements of who were the winners of the AWS Chatbot Challenge as much as I was. Well wait no longer, the winners of the AWS Chatbot Challenge have been decided.

May I have the Envelope Please? (The Trumpets sound)

The winners of the AWS Chatbot Challenge are:

  • First Place: BuildFax Counts by Joe Emison
  • Second Place: Hubsy by Andrew Riess, Andrew Puch, and John Wetzel
  • Third Place: PFMBot by Benny Leong and his team from MoneyLion.
  • Large Organization Winner: ADP Payroll Innovation Bot by Eric Liu, Jiaxing Yan, and Fan Yang

 

Diving into the Winning Chatbot Projects

Let’s take a walkthrough of the details for each of the winning projects to get a view of what made these chatbots distinctive, as well as, learn more about the technologies used to implement the chatbot solution.

 

BuildFax Counts by Joe Emison

The BuildFax Counts bot was created as a real solution for the BuildFax company to decrease the amount the time that sales and marketing teams can get answers on permits or properties with permits meet certain criteria.

BuildFax, a company co-founded by bot developer Joe Emison, has the only national database of building permits, which updates data from approximately half of the United States on a monthly basis. In order to accommodate the many requests that come in from the sales and marketing team regarding permit information, BuildFax has a technical sales support team that fulfills these requests sent to a ticketing system by manually writing SQL queries that run across the shards of the BuildFax databases. Since there are a large number of requests received by the internal sales support team and due to the manual nature of setting up the queries, it may take several days for getting the sales and marketing teams to receive an answer.

The BuildFax Counts chatbot solves this problem by taking the permit inquiry that would normally be sent into a ticket from the sales and marketing team, as input from Slack to the chatbot. Once the inquiry is submitted into Slack, a query executes and the inquiry results are returned immediately.

Joe built this solution by first creating a nightly export of the data in their BuildFax MySQL RDS database to CSV files that are stored in Amazon S3. From the exported CSV files, an Amazon Athena table was created in order to run quick and efficient queries on the data. He then used Amazon Lex to create a bot to handle the common questions and criteria that may be asked by the sales and marketing teams when seeking data from the BuildFax database by modeling the language used from the BuildFax ticketing system. He added several different sample utterances and slot types; both custom and Lex provided, in order to correctly parse every question and criteria combination that could be received from an inquiry.  Using Lambda, Joe created a Javascript Lambda function that receives information from the Lex intent and used it to build a SQL statement that runs against the aforementioned Athena database using the AWS SDK for JavaScript in Node.js library to return inquiry count result and SQL statement used.

The BuildFax Counts bot is used today for the BuildFax sales and marketing team to get back data on inquiries immediately that previously took up to a week to receive results.

Not only is BuildFax Counts bot our 1st place winner and wonderful solution, but its creator, Joe Emison, is a great guy.  Joe has opted to donate his prize; the $5,000 cash, the $2,500 in AWS Credits, and one re:Invent ticket to the Black Girls Code organization. I must say, you rock Joe for helping these kids get access and exposure to technology.

 

Hubsy by Andrew Riess, Andrew Puch, and John Wetzel

Hubsy bot was created to redefine and personalize the way users traditionally manage their HubSpot account. HubSpot is a SaaS system providing marketing, sales, and CRM software. Hubsy allows users of HubSpot to create engagements and log engagements with customers, provide sales teams with deals status, and retrieves client contact information quickly. Hubsy uses Amazon Lex’s conversational interface to execute commands from the HubSpot API so that users can gain insights, store and retrieve data, and manage tasks directly from Facebook, Slack, or Alexa.

In order to implement the Hubsy chatbot, Andrew and the team members used AWS Lambda to create a Lambda function with Node.js to parse the users request and call the HubSpot API, which will fulfill the initial request or return back to the user asking for more information. Terraform was used to automatically setup and update Lambda, CloudWatch logs, as well as, IAM profiles. Amazon Lex was used to build the conversational piece of the bot, which creates the utterances that a person on a sales team would likely say when seeking information from HubSpot. To integrate with Alexa, the Amazon Alexa skill builder was used to create an Alexa skill which was tested on an Echo Dot. Cloudwatch Logs are used to log the Lambda function information to CloudWatch in order to debug different parts of the Lex intents. In order to validate the code before the Terraform deployment, ESLint was additionally used to ensure the code was linted and proper development standards were followed.

 

PFMBot by Benny Leong and his team from MoneyLion

PFMBot, Personal Finance Management Bot,  is a bot to be used with the MoneyLion finance group which offers customers online financial products; loans, credit monitoring, and free credit score service to improve the financial health of their customers. Once a user signs up an account on the MoneyLion app or website, the user has the option to link their bank accounts with the MoneyLion APIs. Once the bank account is linked to the APIs, the user will be able to login to their MoneyLion account and start having a conversation with the PFMBot based on their bank account information.

The PFMBot UI has a web interface built with using Javascript integration. The chatbot was created using Amazon Lex to build utterances based on the possible inquiries about the user’s MoneyLion bank account. PFMBot uses the Lex built-in AMAZON slots and parsed and converted the values from the built-in slots to pass to AWS Lambda. The AWS Lambda functions interacting with Amazon Lex are Java-based Lambda functions which call the MoneyLion Java-based internal APIs running on Spring Boot. These APIs obtain account data and related bank account information from the MoneyLion MySQL Database.

 

ADP Payroll Innovation Bot by Eric Liu, Jiaxing Yan, and Fan Yang

ADP PI (Payroll Innovation) bot is designed to help employees of ADP customers easily review their own payroll details and compare different payroll data by just asking the bot for results. The ADP PI Bot additionally offers issue reporting functionality for employees to report payroll issues and aids HR managers in quickly receiving and organizing any reported payroll issues.

The ADP Payroll Innovation bot is an ecosystem for the ADP payroll consisting of two chatbots, which includes ADP PI Bot for external clients (employees and HR managers), and ADP PI DevOps Bot for internal ADP DevOps team.


The architecture for the ADP PI DevOps bot is different architecture from the ADP PI bot shown above as it is deployed internally to ADP. The ADP PI DevOps bot allows input from both Slack and Alexa. When input comes into Slack, Slack sends the request to Lex for it to process the utterance. Lex then calls the Lambda backend, which obtains ADP data sitting in the ADP VPC running within an Amazon VPC. When input comes in from Alexa, a Lambda function is called that also obtains data from the ADP VPC running on AWS.

The architecture for the ADP PI bot consists of users entering in requests and/or entering issues via Slack. When requests/issues are entered via Slack, the Slack APIs communicate via Amazon API Gateway to AWS Lambda. The Lambda function either writes data into one of the Amazon DynamoDB databases for recording issues and/or sending issues or it sends the request to Lex. When sending issues, DynamoDB integrates with Trello to keep HR Managers abreast of the escalated issues. Once the request data is sent from Lambda to Lex, Lex processes the utterance and calls another Lambda function that integrates with the ADP API and it calls ADP data from within the ADP VPC, which runs on Amazon Virtual Private Cloud (VPC).

Python and Node.js were the chosen languages for the development of the bots.

The ADP PI bot ecosystem has the following functional groupings:

Employee Functionality

  • Summarize Payrolls
  • Compare Payrolls
  • Escalate Issues
  • Evolve PI Bot

HR Manager Functionality

  • Bot Management
  • Audit and Feedback

DevOps Functionality

  • Reduce call volume in service centers (ADP PI Bot).
  • Track issues and generate reports (ADP PI Bot).
  • Monitor jobs for various environment (ADP PI DevOps Bot)
  • View job dashboards (ADP PI DevOps Bot)
  • Query job details (ADP PI DevOps Bot)

 

Summary

Let’s all wish all the winners of the AWS Chatbot Challenge hearty congratulations on their excellent projects.

You can review more details on the winning projects, as well as, all of the submissions to the AWS Chatbot Challenge at: https://awschatbot2017.devpost.com/submissions. If you are curious on the details of Chatbot challenge contest including resources, rules, prizes, and judges, you can review the original challenge website here:  https://awschatbot2017.devpost.com/.

Hopefully, you are just as inspired as I am to build your own chatbot using Lex and Lambda. For more information, take a look at the Amazon Lex developer guide or the AWS AI blog on Building Better Bots Using Amazon Lex (Part 1)

Chat with you soon!

Tara