All the Major Announcements from AWS re:Invent 2019

If you asked a cloud architect just six years ago how much they knew about AWS, the reply would be optimistically 70 percent. Then re:Invent would come along, new products and services would be announced, and if you asked the same architect the same question after the conference, the answer would be a lot less. 

This year's re:Invent was no exception. Over 65,000 people from around the globe and industry verticals converged on the massive conference campus in Las Vegas. Some came to learn, network, exchange ideas, converse with industry pros, recruit, collect swag and most importantly, to get a glimpse of the new products and features that are unveiled during the course of the conference. It's a lot like an Apple product event, if the event went on for a week with potentially paradigm shifting announcements every day.

After attending a handful of these, patterns start to emerge on what gets announced. At a high-level, announcements fall into two categories:

• Disruptive – fundamentally altering how something has always been done
• Evolutionary – changes to core services to provide more value to customers

AWS announced 77 new product launches, feature releases and services at this year's re:Invent conference, and that number is even higher if you count all the releases leading up to the conference. Obviously, if we tried to do a recap of every single announcement, it would be approximately twenty pages long and make fantastic indexing material, but nothing would stick. Instead, let's touch upon the significant disruptive and evolutionary announcements this year and expound a bit on why they are important.

AWS Outposts

Technically announced last year, AWS Outposts was made GA during this re:Invent with some added features hinted at during the previous announcement. Put simply, it's AWS on-prem (either EC2 or VMC flavored). Outposts can be ordered from the AWS console and assuming all the prerequisites are met, shipped and installed in a customer's data center, allowing them to automate/provision/manage workloads on the Outpost as if they were in the cloud. Services supported at launch include EC2, EBS, ECS, EKS, RDS, EMR and AppMesh.

Why it's important:

A lot of our customers have data sovereignty concerns and want to keep their data on-prem. While solutions such as utilizing colocations exists, that can itself be challenging to architect. Outposts provides the flexibility of cloud compute with low latency access to data sitting on-prem. 

AWS Local Zones and AWS Wavelength

Several new launches focused on bringing AWS closer to the edge and the data center, such as AWS Local Zones, which brings AWS infrastructure closer to customers that require single digit millisecond latencies for local end users. This is a benefit of AWS Wavelength as well, which brings AWS compute and storage to the telecommunication providers' 5G edge.

Why it's important:

This is very exciting as one of the main challenges with our ever-growing connectivity is latency, which has been a barrier for emerging technologies such as autonomous machinery and IoT.

EKS for Fargate

An evolutionary iteration for the popular Fargate service, EKS for Fargate allows customers to run their Kubernetes Nodes on Fargate, eliminating the need to maintain EC2 infrastructure for EKS cluster nodes. 

Why it's important:

AWS wants to make it easy for customers to get up and running with Kubernetes in the cloud without having to worry about maintaining the infrastructure around it, which can become very complex very quickly. Utilizing this service allows customers to focus more on applications than infrastructure. The built-in right-sizing of Fargate can also drive down costs of infrastructure. Combined with two other announcements (Fargate Spot and Savings Plan), customers can utilize the benefits of Fargate and save money while doing it.

SageMaker announcements

AWS made many evolutionary announcements to expand and round out its SageMaker service — so much that we won't go into each announcement here. However, they did add notable additions such as SageMaker Studio, an IDE for Data Scientists and Machine Learning Specialists, SageMaker Experiments, SageMaker Debugger, SageMaker AutoPilot and SageMaker Operator for Kubernetes, to name a few. 

Why it's important:

We are moving more and more towards a data driven world and experience. Many organizations are turning to the power and flexibility of the cloud to build, test, train and deploy their machine learning models, and AWS wants to make this as easy as possible for today's data scientists. 

Amazon CodeGuru

Another large focus area and theme of this year's conference was the ability to make AI/ML more accessible and usable to those without the background. CodeGuru helps developers by using ML to evaluate code, as well as monitor performance of code and provide recommendations for improvements.

Why it's important:

Trained against thousands of internal and open-source projects, CodeGuru can identify common problem areas such as resource leaks and race conditions, and additionally assist in optimizing code for performance.  Currently CodeGuru only supports Java, but AWS has plans to add languages to this service in the future.

Amazon Detective

Security in the cloud is of paramount concern for both customers and cloud providers. A myriad of products exists, both native and from partners to help enhance security and provide visibility into the security posture of an environment. Amazon Detective takes in log data from customer AWS resources and runs it through machine learning algorithms, statistical analysis and graph theory to help customers identify root causes to problems in their cloud environments. 

Why it's important:

Collecting all the various logs and data from so many resources can prove very daunting to digest and determine linkages, and having a tool that establishes those connections for you allows for quick analysis of causality.

AWS Transit Gateway Network Manager & networking announcements

While not as big of a shift as the Transit Gateway announcement last year (thousands of Transit VPCs were snapped out of existence), a flurry of augments to TGW were announced this year that will have to be accounted for when designing network architectures in the cloud: 

• AWS Transit Gateway Network Manager  provides a single global view to manage your network across AWS and on premise
• The ability to assign route tables to edge constructs in a VPC such as Internet Gateways and Virtual Gateways allows customers the ability to route incoming and outcoming traffic from the edge of the VPC and simplify such things as a "centralized ingress VPC" used for traffic inspection.
• Transit Gateway Inter-Region peering allows multiple Transit Gateways to be peered together to create global networks more easily and efficiently. Also, the Inter-Region peering can be used to transitively route from a Transit Gateway with no Transit VIF attachment back to one that does, to connect multiple regions back to an on-premise location.
• Accelerated Site-to-Site VPN gives Transit Gateway VPN connections a boost (whether you want to connect back to a data center over a Direct Connect or VPN), utilizing AWS Global Accelerator to provide better reliability and latency to your private resources.
• Transit Gateway support for Multicast Routing is now possible in AWS without complex tunneling overlay workarounds.

Why it's important:

For a deeper analysis of the network announcements, stay tuned for another upcoming article focusing on the network considerations at re:Invent 2019.

Amazon Braket

The last announcement to be covered would definitely fall under the disruptive category. Amazon Braket is a fully managed service that allows customers to build, test and run quantum computing applications against different quantum computers from AWS partners. 

Essentially AWS is providing an environment that allows users to explore quantum computing across different hardware implementations. Quantum computing is an area that has fascinated computer scientists on its possibilities for some time, but the technology to bring it to fruition has remained elusive. Now, with the power of cloud titans, we may be closer to that age.

Why it's important:

Because it's a window into the future.  

What's next?

At the end of every re:Invent, we feel a case of information overload. While we've highlighted just a handful of new offerings and features unveiled during the conference that we felt were important, there's a lot more that were left out for the sake of brevity. However, what is important to you will depend on your role, interests and expertise.

Stay up to date with the latest from AWS and more by creating an account and exploring our comprehensive lab offerings.