top of page

re:Invent 2019 - Announcements Highlights



Hey everyone, some important updates from re:Invent!


This is a partial list and will update as re:Invent announcements are made.


EKS on Fargate is now GA:

https://aws.amazon.com/blogs/aws/amazon-eks-on-aws-fargate-now-generally-available/

EKS on Fargate allows us to create pods in Fargate without the need to take care of the underlying nodes.

Scale-out and scale-in is not required.

We still have to pay the 0.20$ / hour for the cluster itself.

Limitations: no privileged pods, not more than 4 vCPU, no persistent workloads (no PVC in Fargate EKS), no DaemonSets, pods can’t use HostNetwork or NodePort, ALB is the only supported LB.


AWS Fargate Spot:

https://aws.amazon.com/blogs/aws/aws-fargate-spot-now-generally-available/

For customers using Fargate for ECS to run interruption tolerant workloads, this feature is a big deal!

ECS spots can lower prices up to 70% in comparison to standard Fargate price.

I hope we'll see it in Fargate for EKS as well, with native integration to Kubernetes Jobs.


Transit Gateway updates:

https://aws.amazon.com/blogs/aws/aws-transit-gateway-adds-multicast-and-inter-regional-peering/

Multicast support for Transit Gateway (and VPC, of course).

AWS now allows you to run native Multicast across multiple VPCs. It can be beneficial for migrating traditional workloads to the cloud.

AWS is now the only major cloud provider that supports multicast.

Multi-region peering on Transit Gateway. Very useful to allow "transit VPC" replacement.

Another cool feature they released is centralized visualization of global networks using Network Manager:

https://aws.amazon.com/blogs/aws/new-for-aws-transit-gateway-build-global-networks-and-centralize-monitoring-using-network-manager/


AWS Compute Optimizer:

https://aws.amazon.com/blogs/aws/aws-compute-optimizer-your-customized-resource-optimization-service/

Uses Machine Learning to analyze past metrics on EC2 workloads to allow better rightsizing than the available options today.


AWS Inferentia custom Machine Learning Inferences chip is now available with new instance size Inf1:

https://aws.amazon.com/blogs/aws/amazon-ec2-update-inf1-instances-with-aws-inferentia-chips-for-high-performance-cost-effective-inferencing/

https://aws.amazon.com/machine-learning/inferentia/

AWS has been working on its ML inferences chip design.

It was announced a year ago, and now it's generally available through the Inf1 instance type.

Pricing is lower than the P3 and P3dn instances, but we'll have to wait a few days for the Amazon Inferentia benchmarks.


AWS Outposts:

https://aws.amazon.com/blogs/aws/aws-outposts-now-available-order-your-racks-today/

AWS has announced it's long-awaited service for running region-like services on-premise or co-located.

It's especially interesting for the Israeli market as it'll allow heavy regulated industries that couldn't go to the cloud to bring the cloud to them.

The control-plane resides in the desired region (for the Israeli market, although not available yet, most likely Frankfurt or Ireland).

Outposts support the following services from day one: EC2, EBS, VPC, ECS, EKS, EMR.

RDS for PostgreSQL and MySQL are available in preview.

More services to come.

At first glance, pricing seems a bit high. But I'll have to get a quote from hardware vendors to see what's the tipping point for Outposts.

https://aws.amazon.com/outposts/pricing/


AWS Local Zone Los Angeles in US-West-2 (Oregon):