AWS Re:Invent 2020 Announcements, Updates and Recap
AWS re:invent conference organized by AWS this year was virtual, free and almost three weeks long. AWS announced new features, improvements and new products and cloud services. Following is a quick review of the important highlights of the conference impacting compute, database, storage, machine learning and Data Analytics. https://reinvent.awsevents.com
Highlights
- S3 Storage Lens — New Dashboard in S3 Console to check for Data Lake Insights and outliers
- EKS and ECS Anywhere — Allows you to run ECS and EKS anywhere in On-premise or any other data center.
- Aurora Serverless V2 — Allows you to quickly scale to as big as you want in fraction of seconds, along with features of read replicas, Multi AZ
- Redshift AQUA — High-speed cache on top of Redshift Managed Storage
Storage
S3 Updates:
- S3 Outposts that allows you to add storage of 48TB or 96TB and S3 Data Sync can be used to keep the data sync between Outpost and AWS.
- S3 Intelligent Tiering is extended to 4 Access Tiers (Standard, Infrequent, Archive and Deep Archive)
- S3 now has Strong Consistency which will be of great benefit when you are running analytics directly on top of Data Lake and you always want to read the latest record
- S3 Replication will support Bi-Directional Replication, Multi Destination Replication and also provides replication notifications with metrics.
- S3 Security Update Features now include Bucket Level Owner Overwrite.
EBS Updates
- EBS — GP3 Volumes which separates out IOPS and Throughput from Storage and optimizes storage cost.
Compute
- EKS Anywhere and ECS Anywhere allows you to run AWS Services anywhere like On-premise or any Data Center.
- Lambda compute cost reduced by 1ms instead of 100ms, also now supports 10GB memory and 6 VCPU’s and now supports Container images as a packaging format.
- AWS CloudShell is a browser-based shell that makes it easy to securely manage, explore, and interact with your AWS resources. CloudShell is pre-authenticated with your console credentials.
Database
- Aurora Serverless V2 — Scales as big as you want to in fraction of seconds, Multi AZ, Read replicas. MySQL version this year, PostgreSQL early next year
- Babelfish for Aurora PostgreSQL — will work with Application code that uses SQL Server to directly work on using Aurora and Babelfish will be made open source
- Redshift AQUA — AQUA is a high-speed cache on top of Redshift Managed Storage that can scale out and process data in parallel across many AQUA nodes.
Machine Learning, IOT and Analytics
- SageMaker — has made significant improvements with the following features
- SageMaker Studio — Integrated tools for SageMaker
- SageMaker Data Wrangler — Comes with 300 built-in transformations for data preparation
- SageMaker Feature Store — Helps store and share machine learning features
- SageMaker Pipelines — Helps run all the steps involved in preparing data and running models
2. AWS Proton — Will be used as a service for management and deployment of containers and Serverless processes.
3. DEVOPS Guru — Makes it easy for Developers to identify operational issues, Uses machine learning to identify the configuration issues
4. AWS Glue Elastic Views — Works like materialized views using different data stores like Dynamo DB, Elastic Search or Redshift
- It can be used for Combine or replicate into a target Data Store
- Any changes in source will be propagated to target data store
5. Quicksight Q — Ask any question in NL and dashboard with results will be displayed
6. AWS Wavelength — Optimized for 5G Mobile Computing