Tag Archives: AWS VPN

Building Resilient and High Performing Cloud-based Applications in Hawaii

2022-01-21 Marie Yap

Post Syndicated from Marie Yap original https://aws.amazon.com/blogs/architecture/building-resilient-and-high-performing-cloud-based-applications-in-hawaii/

Hawaii is building a digital economy for a sustainable future. Many local businesses are already embarking on their journey to the cloud to meet their customers’ growing demand for digital services. To access Amazon Web Services (AWS) on the US mainland, customers’ data must traverse through submarine fiber-optic cable networks approximately 2,800 miles across the Pacific Ocean. As a result, organizations have two primary concerns:

Resiliency concerns about multiple outage events that could arise from breaks in the submarine cables.
Latency concerns for mission-critical applications driven by physical distance.

These problems can be solved by architecting the workloads for reliability, secure connectivity, and high performance.

Designing network connectivity that is reliable, secure, and highly performant

A typical workload in AWS can be broken down into three layers – Network, Infrastructure, and Application. For each layer, we can design for resiliency and latency concerns. Starting at the network layer, there are two recommended options for connecting the on-premises network within the island to AWS.

Use of AWS Direct Connect over a physical connection. AWS Direct Connect is a dedicated network connection that connects your on-premises environment to AWS Regions. In this case, the connection is traversing the fiber-optic cable across the Pacific Ocean to the mainland’s meet-me-point facilities. It can be provisioned from 50 Mbps up to 100 Gbps. This provides you with a presence in an AWS Direct Connect location, a third-party colocation facility, or an Internet Service Provider (ISP) that provides last-mile connectivity to AWS. In addition, the Direct Connect location establishes dedicated connectivity to Amazon Virtual Private Clouds (VPC). This improves application performance and addresses latency concerns by connecting directly to AWS and bypassing the public internet.
Use of AWS VPN over an internet connection. As a secondary option to Direct Connect, AWS Site-to-Site VPN provide connectivity into AWS over the public internet using VPN encryption technologies. The Site-to-Site VPN connects on-premises sites to AWS resources in an Amazon VPC. As a result, you can securely connect your on-premises network to AWS using an internet connection.

We recommend choosing the us-west-2 AWS Region in Oregon to build high performant connectivity closest to Hawaii. The us-west-2 Region generally provides more AWS services at a lower cost versus us-west-1. In addition, there are various options for AWS Direct Connect Locations in the US West Region. Many of these locations support up to 100 Gbps and support MACsec, which is an IEEE standard for security encryption in wired Ethernet LANs. Typically, customers will use multiple 10-Gbps connections for higher throughput and redundancy.

Subsea Cable	Hawaii Cable Landing Station	Mainland Cable Landing Station	Nearest Direct Connect Location
Southern Cross Cable Network (SCCN)	Kahe Point (Oahu)	Morro Bay, CA	CoreSite, Equinix
Southern Cross Cable Network (SCCN)	Kahe Point (Oahu)	Hillsboro, OR	Equnix, EdgeConnex, Pittock Block, CoreSite, T5, TierPoint
Hawaiki	Kapolei (Oahu)	Hillsboro, OR	Equnix, EdgeConnex, Pittock Block, CoreSite, T5, TierPoint
Asia-America Gateway (AAG)	Keawaula (Oahu)	San Luis Obispo, CA	CoreSite, Equinix
Japan-US Cable Network (JUS)	Makaha (Oahu)	Morro Bay, CA	CoreSite, Equinix
SEA-US	Makaha (Oahu)	Hermosa Beach, CA	CoreSite, Equinix, T5

Table 1. Subsea fiber-optic cables connecting Hawaii to the US mainland

(Source: Submarine Cable Map from TeleGeography)

To build resilient connectivity, six cables connect Hawaii to the mainland US: Hawaiki, SEA-US, Asia-America Gateway (AAG), Japan-US (JUS), and two Southern Cross (SCCN) cables. In addition, these cables connect to various locations on the US West Coast. If you require high resiliency, we recommend a minimum of two physically redundant Direct Connect connections into AWS. In addition, we recommend designing four Direct Connect connections that span two Direct Connect locations for maximum resiliency. If you build your architecture following these recommendations, AWS offers this published service level agreement (SLA).

Figure 1. Redundant direct connection from Hawaii to the US mainland

Most customers select an ISP to get them connectivity across the Pacific Ocean to an AWS Direct location. The Direct Connect locations are third-party colocation providers who act as meet-me points for AWS customers and the AWS Regions. For example, our local AWS Partner DRFortress connects multiple ISPs in a data center in Hawaii to the AWS US West Region. We recommend having at least two ISPs for resilient applications, each providing connectivity across a separate subsea cable from Hawaii to the mainland. If one cable should fail for any reason, connectivity to AWS would still be available. The red links in figure 2 are the ISP-provided connectivity that spans the Pacific Ocean. This is a minimum starting point for business-critical applications and should be designed with additional Direct Connect links for greater resiliency.

Architecting for high performance and resiliency

Moving from the network to the infrastructure and application layer, organizations have the option in building their application all in the cloud or in combination with an on-premises environment. An example of an application built all in the cloud is the LumiSight platform in AWS built by local AWS Partner, DataHouse. LumiSight has helped dozens of organizations quickly and securely reopen safely during the pandemic.

Other customers need a hybrid cloud architecture solution. These organizations require that their data processing and locally hosted applications analysis is close to other components within the island’s data center. With this proximity, they can deliver near real-time responses to their end users. AWS Outposts Family extends the capabilities of an AWS Region to the island. This enables local businesses to build and run low latency applications on-premises on an AWS fully managed infrastructure. You can now deploy Compute, Storage, Containers, Data Analytics clusters, Relational, and Cache databases in high performance, redundant and secure infrastructure maintained by AWS. Outposts can be shipped to Hawaii, connecting to the us-west-1 or us-west-2 Regions.

Another option for improving application performance is providing an efficient virtual desktop to access their applications anywhere. Amazon WorkSpaces provides a secure, managed cloud-based virtual desktop experience. Many workers who bring their own device (BYOD) or work remotely use Workspaces to access their corporate applications securely. Workspaces use streaming protocols that provide a secure and responsive desktop experience to end users located in remote Regions, like Hawaii. Workspaces can quickly provide a virtual desktop without managing the infrastructure, OS versions, and patches. You can test your connection to Workspaces from Hawaii, or anywhere else in the world, at the Connection Health Check page.

Architecting for resiliency in the infrastructure and application stack is vital for Business Continuity and Disaster Recovery (BCDR) plans. Organizations in Hawaii who are already using VMware can take advantage of creating a recovery site using VMware Cloud on AWS as their solution for disaster recovery. The VMware Cloud on AWS is a fully managed VMware software-defined Data Center (SDDC) running on AWS, which provides access to native AWS services. Organizations can pair their on-premises vCenter and virtual machines to the fully managed vCenter and virtual machines residing in the cloud. The active Site Recovery Manager provides the automation of failing over and failing back applications between on-premises to the cloud DR site and vice versa. Additionally, organizations can define their SDDC in the us-west-2 Region using AWS Direct Connect to minimize the latency of replicating the data from and to the data centers in the islands.

Conclusion

Organizations in Hawaii can build resilient and high performant cloud-based workloads with the help of AWS services in each layer of their workloads. Starting with the network layer, you can establish reliable and lower latency connectivity through redundant AWS Direct Connect connections. Next, for low latency, hybrid applications, we extend infrastructure capabilities locally through AWS Outposts. We also improve the user experience in accessing cloud-based applications by providing Amazon WorkSpaces as the virtual desktop. Finally, we build resilient infrastructure and applications using a familiar solution called VMware Cloud on AWS.

To start learning the fundamentals and building on AWS, visit the Getting Started Resource Center.

Configuring AWS VPN for UK public sector use

2020-12-02 Charlie Llewellyn

Post Syndicated from Charlie Llewellyn original https://aws.amazon.com/blogs/security/configuring-aws-vpn-for-uk-public-sector-use/

In this post, we explain the United Kingdom (UK) National Cyber Security Centre (NCSC)’s guidance on VPN profiles configuration, and how the configuration parameters for the AWS Virtual Private Network (AWS VPN) align with the NCSC guidance. At the end of the post, there are links to code to deploy the AWS VPN in line with those parameters.

Many public sector organizations in the UK need to connect their existing on-premises facilities, data centers, or offices to the Amazon Web Services (AWS) cloud so they can take advantage of the broad set of services AWS provides to help them deliver against their mission.

This can be achieved using the AWS VPN service. However some customers find it difficult to know the exact configuration parameters that they should choose when establishing the VPN connection in-line with guidance for the UK public sector.

AWS VPN services enable organizations to establish secure connections between their on-premises networks, remote offices, and client devices and the AWS global network. AWS VPN comprises two services: AWS Site-to-Site VPN and AWS Client VPN. Together, they deliver a highly available, fully managed, elastic cloud VPN solution to protect your network traffic.

For the purposes of this post, we focus on the Site-to-Site VPN configuration, not Client VPN because the NCSC guidance we’re discussing is specifically related to site-to-site VPNs. This post covers two areas:

An overview of the current guidance for VPN configurations for the public sector.
Recommendations on how to configure AWS VPN to meet or exceed the current guidance.

VPN guidance for UK public sector organizations

The starting point for security guidance for the UK public sector is often the NCSC. The role of the NCSC includes:

Protecting government systems and information.
Planning for and responding to cyber incidents.
Working with providers of critical national infrastructure to improve the protection and computer security of such infrastructure against cyber-borne threats.

Specifically, for guidance on the configuration of VPNs for the UK public sector to support data at OFFICIAL, the NCSC has created detailed guidance on the technical configurations to support two different profiles: PRIME and Foundation. These two profiles provide different technical implementations to support different equipment and are both suitable for use with OFFCICIAL data. Beyond these technical differences, NCSC also documents that Foundation is expected to provide suitable protection for OFFICIAL information until at least December 31, 2023, while PRIME has no review date specified at the time of writing.

This guidance is available in Using IPsec to protect data.

Let’s start by debunking a few myths.

Myth 1: I have to adhere exactly to the NCSC technical configuration or I cannot use a VPN for OFFICIAL data

It’s a common misconception that a public sector organization must adhere exactly to the configuration of either PRIME or Foundation in order to use a VPN for OFFICIAL data, even if other configuration options available—such as a longer key length—offer a higher security baseline.

Note that the NCSC isn’t mandating the use of the configuration in their guidance. They’re offering a configuration that provides a useful baseline, but you must assess your use of the NCSC guidance in context of the risks. To help with these risk-based decisions, the NCSC has developed a series of guidance documents to help organizations make risk-based decisions. A common consideration that might require deviating from the guidance would be supporting interoperability with legacy systems where the suggested algorithms aren’t supported. In this case, a risk-based decision should be made—including accounting for other factors such as cost.

It’s also worth noting that the NCSC creates guidance designed to be useful to as many organizations as possible. The NCSC balances adopting the latest possible configurations with backwards compatibility and vendor support. For example, the NCSC suggests AES-128 where—in theory—AES-256 could also be a good choice. Organizations need to be aware that if they choose to adopt devices that support only AES-256 and later need to connect in devices capable of only AES-128, there could be significant investment to replace the legacy devices with ones that support AES-256. However, AWS provides both AES-128 and AES-256, so if the remote device supports it, AWS would recommend opting for AES-256.

The NCSC also tries to develop advice that has some longevity. For example, the guidance suggesting use of AES-128 was created in 2012 with a view to providing solid guidance over a number of years. This means customers can choose different configuration parameters that offer increased levels of security if both sides of the VPN can support it.

It’s possible for a customer to choose options that might lower the security of the connection, provided that risks are identified and appropriately managed by the customers assurance team. This might be needed to support interoperability between existing systems where the cost of an upgrade outweighs the risk.

Myth 2: Foundation has been deprecated and I must use PRIME

Another common misconception is that Foundation has been deprecated in favor of PRIME. This is not the case. The NCSC has stated that Foundation is expected to provide suitable protection for OFFICIAL information until at least December 31, 2023. The security provided by both solutions provides commensurate security for accessing data classified as OFFICIAL. One of the main differences between PRIME and Foundation is the choice of signature algorithm: RSA or ECDSA. This difference can be helpful in enabling an organization to choose which profile to adopt. For example, if the organization already has a private key infrastructure (PKI), then the decision regarding which signature algorithm to use is based on what existing systems support.

Myth 3: I can’t use Foundation for accessing OFFICIAL SENSITIVE data

A final point that often causes confusion is the classifying of data at OFFICIAL SENSITIVE because it isn’t a classification, but a handling caveat. The data would be classified as OFFICIAL and marked as OFFICIAL SENSITIVE, meaning that systems handling the data need risk-appropriate security measures. A system that can handle OFFICIAL data might be appropriate to handle sensitive information. Hence Foundation could be suitable for accessing OFFICIAL SENSITIVE data, depending on the risks identified.

Deep-dive into the technical specifications

Now that you know a little more about how the guidance should be viewed, let’s look more closely at the technical configurations for each VPN profile.

The following table shows the configuration parameters suggested by the NCSC VPN guidance discussed previously.

Technical detail	Foundation	PRIME
IKEv* – Encryption	IKEv1 – AES with 128-bit keys in CBC mode (RFC3602)	IKEv2 – AES-128 in GCM-128 (and optionally CBC)
IKEv2 – Pseudo-random function	HMAC-SHA256	HMAC-SHA256
IKEv2 – Diffie-Hellman group	Group 14 (2048-bit MODP group) (RFC3526)	256 bit random ECP (RFC5903) Group 19
IKEv2 – Authentication	X.509 certificates with RSA signatures (2048 bits) and SHA-256 (RFC4945 and RFC4055)	X.509 certificates with ECDSA-256 with SHA256 on P-256 curve
ESP – Encryption	AES with 128-bit keys in CBC mode (RFC3602) SHA-256 (RFC4868)	AES-128 in GCM-128 SHA-256 (RFC4868)

Recommended AWS VPN configuration for public sector

Bearing in mind these policies, and remembering that the configuration is only guidance, you must make a risk-based decision. AWS recommends the following configuration as a starting point for the configuration of the AWS VPN.

Technical detail	AWS configuration	Adherence
IKEv* – Encryption	IKEv2 – AES-256-GCM	Suitable for Foundation and PRIME
IKEv2 – Pseudo-random function	HMAC-SHA256	Meets Foundation and PRIME
IKEv2 – Diffie-Hellman group	Group 19	Suitable for Foundation and matches PRIME
IKEv2 – Authentication	RSA 2048 SHA2-512	Suitable for Foundation
ESP – Encryption	AES-256-GCM	Suitable for PRIME and Foundation

In the table above, we use the term suitable for where the protocol doesn’t match the guidance exactly but the AWS configuration options provide equivalent or stronger security—for example, by using a longer key length.

With the configuration defined above, the AWS VPN service is suitable for use under the Foundation profile in all areas. It can also be made suitable for PRIME in all areas apart from IKEv2 encryption. The use of RSA or ECDSA is the main difference between the AWS VPN and PRIME configurations. This makes the current AWS VPN solution closer to Foundation than PRIME.

When considering which options are available to you, the starting point should be the capabilities of your current—and possible future—VPN devices. Based on its capabilities, you can use the NCSC guidance and preceding tables to choose the protocols that match or are suitable for the NCSC guidance.

Summary

To review:

The NCSC provides guidance for the VPN configuration, not a mandate.
An organization is free to decide not to use the guidance, but should consider risks when they make that decision.
The AWS VPN meets or is suitable for the configuration options for Foundation.

After reviewing the details contained in this blog, UK public sector organizations should have the confidence to use the AWS VPN service with systems running at OFFICIAL.

If you’re interested in deploying the AWS VPN configuration described in this post, you can download instructions and AWS CloudFormation templates to configure the AWS VPN service. The AWS VPN configuration can be deployed to either connect directly to a single Amazon Virtual Private Cloud (Amazon VPC) using a virtual private gateway, or to an AWS Transit Gateway to enable its use by multiple VPCs.

If you’re interested in configuring your AWS VPN tunnel options manually, you can follow Modifying Site-to-Site VPN tunnel options.

If you have feedback about this post, submit comments in the Comments section below.

Want more AWS Security how-to content, news, and feature announcements? Follow us on Twitter.

New Whitepaper: Selecting & Designing Your Hybrid Connectivity Model

2020-10-08 Santiago Freitas

Post Syndicated from Santiago Freitas original https://aws.amazon.com/blogs/architecture/new-whitepaper-selecting-designing-your-hybrid-connectivity-model/

Introduction

Many organizations need to connect their on-premises data centers, remote sites, and the cloud. A hybrid network connects these different environments.

A modern organization uses an extensive array of IT resources. In the past, it was common to host these resources in an on-premises data center or a colocation facility. With the increased adoption of cloud computing, IT resources are delivered and consumed from cloud service providers over a network connection. In some cases, organizations have opted to migrate all existing IT resources to the cloud. In other cases, organizations maintain IT resources both on premises and in the cloud. In both cases, a common network is required to connect on-premises and cloud resources. Coexistence of on-premises and cloud resources is called “hybrid cloud” and the common network connecting them is referred to as a “hybrid network. “ Even if your organization keeps all of its IT resources in the cloud, it may still require hybrid connectivity to remote sites.

There are several connectivity models to choose from. Although having options adds flexibility, selecting the best option requires analysis of the business and technical requirements and the elimination of options that are not suitable. Requirements can be grouped together across considerations, such as: security, time to deploy, performance, reliability, communication model, scalability, and more. Once requirements are carefully collected, analyzed, and considered, network and cloud architects identify applicable AWS hybrid network building blocks and solutions. To identify and select the optimal model(s), architects must understand advantages and disadvantages of each model. There are also technical limitations that might cause an otherwise good model to be excluded.

Figure 1 – Consideration covered on the whitepaper.

A new whitepaper on Hybrid Connectivity describes AWS building blocks and the key things to consider when deciding which hybrid connectivity model is right for you. To help you determine the best solution for your business and technical requirements, we provide decision trees to guide you through the logical selection process as well as a customer use case to show how to apply the considerations and decision trees in practice.

Decision tree applied to Example Corp. Automotive use case

Figure 2: Example Corp. Automotive connection type decision tree

Contributors

Contributors to this new whitepaper on Hybrid Connectivity are: Marwan Al Shawi, AWS Solutions Architect; Santiago Freitas, AWS Head of Technology; Evgeny Vaganov, AWS Specialist Solutions Architect – Networking; and Tom Adamski, AWS Specialist Solutions Architect – Networking. Special thanks to Stephen Bird, AWS Senior Program Manager – Content.

Noise