This year ‘s annual, free and three – week re: invent conference was organized by AWS. Through many keynote speeches and sessions, AWS announced new features, enhancements and cloud services. Below is a review of the key designations affecting computing, database, storage, networking, machine learning and development.
Compute
On the first day of the conference, Amazon announced EC2 Mac examples for macOS, adding a new operating system to EC2 after several years. This is mainly focused on processes that only run on Mac OS, such as building and testing applications for iOS, MacOS, tvOS and Safari. The first part of Andy Jassy’s keynote speech focused on notifications related to computing options and serverless technologies. AWS introduced new sample types on various EC2 processors and families, including Intel Xeon M5zn specs, Graviton2-powered C6gn instances, Intel-powered D3 / D3en specs, full-memory R5b instances and G4ad GPU specs with AMD power. Watch the InfoQ broadcast here.
Lambda reports and use without a server: billing granularity is reduced from 100ms to 1ms which automatically reduces costs for all Lambda operations, tasks with up to 10 GB of memory and 6 vCPU were added ris. Another new feature is the support for container images as a packaging format, to simplify the transition from vessel-based workload to non-server tasks. To learn more about AWda Lambda updates, check out this InfoQ article.
Installing Amazon ECS Anywhere and Amazon EKS Anywhere, AWS makes the container orchestration software used in ECS and EKS freely available for use outside of AWS, providing including other cloud providers. This will provide more integration and lower latency and will follow the path of Microsoft and Google who already offer Azure AKS and Google Anthos for free.
During the first keynote, the public preview of AWS Proton, a new management deployment service for container and serverless applications, was announced. With AWS Proton, customers can automate and manage infrastructure provisioning and code deployment for non-server and ship-based applications. Watch the InfoQ broadcast here. ECR Public Repositories is a public register for the storage, management, sharing and use of images worldwide.
Storage
Three key announcements for EBS, the blockchain storage service designed for use with EC2, included a new Ep measurement type gp3 that is 20% cheaper than the previous gp2 type. In addition, it provides higher baseline performance and is the first general purpose book that allows IOPS to be configured independently of disk size. With attention to the improvements and simplicity of the update, Corey Quinn, cloud economist at The Duckbill Group, proposing immediate switch to the new size:
EBS gp3 is a game changer, a stop-point. 80% of the cost of gp2, can be converted instead, and there is no downside. Do it immediately.
A new type io2 Block Express is available in preview, aided by a small-size IOPS workload, and linear IOPS prices for the io2 type of measure are now in place.
The main update for item storage was that S3 now automatically delivers strong read-write consistency for all applications. AWS announced other enhancements to S3, including multi-destination replication, two-way replication across regions to improve support for multi-master and multi-area applications and new bucket keys .
Database
Important additions to databases included the Babelfish previews for Aurora, a translation cover for Amazon Aurora PostgreSQL that allows Aurora to understand commands from applications written for Microsoft SQL Server. Aurora Serverless v2 is a new server-free, MySQL-compatible relationship database, and AWS Glue Elastic Views builds valuable views that combine and reproduce data across multiple databases. There are separate configurations for both Aurora Serverless v2 and Babelfish for Aurora on InfoQ. Amazon Aurora PostgreSQL is now merging with AWS Lambda.
Various new enhancements and features have been introduced for Amazon Redshift, the database service, for example the ability to move transfers between access zones, automated table optimization and data sharing preview and support for data processing Native JSON.
Networking and IoT
AWS Local Zones, a single zone extension of regions adjacent to densely populated areas to provide lower latency, were introduced in 2019. During the conference, AWS announced that there will be three new Local Zones GA this month in Boston, Houston, and Miami and a dozen more in 2021, including New York City and Chicago.
Smaller Outpost options will continue next year as well, allowing customers to install AWS hardware in smaller offices, factories, and sites that are limited by space but need access to low latency computing capacity.
In place of IoT, the new AWS IoT Greengrass 2.0 provides an open running range and tools for local software development and management on large device fleets.
Machine Learning
Many of the new features and results of machine learning were discussed at the keynote address of Swami Sivasubramanian, vice president of AI and machine learning at AWS, and focused on SageMaker. The new Amazon SageMaker Feature Store is a fully managed store, built specifically to store, update, retrieve and share machine learning features. Watch the InfoQ broadcast here.
Other new services and features mentioned were SrifMaker Clarify, SageMaker Debugger, SageMaker Managed Data Parallelism and SageMaker Model Parallelism. Among many positive reactions, Jeremy Edberg, CEO at MinOps, highlights the benefits of Amazon SageMaker Clarify, the service focusing on bias detection and explanation:
It will help you to detect biases in your dataset. I think it’s wonderful; it just comes to the surface that this problem exists. Many people do not even recognize that this is a problem at all. It’s absolutely amazing!
Coney Quinn’s new approach is confusing:
At its launch, Amazon SageMaker was an easy way to learn tools for people without formal data science training. Today, SageMaker Autopilot, SageMaker Studio, SageMaker Feature Store, SageMaker DataWrangler, SageMaker Ground Truth, SageMaker Notebook, SageMaker Neo, SageMaker RL, SageMaker Marketplace, SageMaker Experiments, SageMaker Debugger, SageMaker Model Monitor, and whatever release between time I write this and the time I publish it means a neophyte is going to pull up the service page, get out of it, shut down the laptop and walk away.
AWS Trainium, an EC2-based machine learning training chip designed by AWS and Habana Gaudi built for machine learning was announced at the start of the conference, along with Amazon Monitron, an end-to-end system that detects my behavior -standard in industrial devices, and Panorama AWS, machine learning tool and SDK.
Amazon Lookout for Metrics is one more add-on in the anomaly detection space, adding a flexible service for time series analysis, along with Amazon Lookout for Equipment and Amazon Lookout for Vision.
Research, Architecture and Coding
An important part of Werner Vogels ’key brand was dedicated to the importance of logging, monitoring and how to improve usage. A few new services and enhancements are now available: CloudTrail provides more granular control over data event logging, while Amazon’s Managed Service for Prometheus (AMP) and Amazon’s Managed Service for Grafting (AMG) now in preview.
AWS Fault Injection Simulator is a chaos-driven engineering service that will be available in 2021 and will include riot events across a range of AWS services, including EC2, EKS, ECS and RDS.
Cloudshell, a browser-based shell for interacting with AWS tools, is already available and makes it simpler to work with the CLI without running a sample and handling credentials. Python is now supported by Amazon CodeGuru, the managed service for automated code reviews and application performance recommendations. Amazon Location, a service to link maps, location sensing, and other location-based features to web and mobile applications was announced in a preview at the end of the conference.
AWS added more: Tactics sessions, for example on S3 topics, from January 12 to January 14, extending the conference to the new year.