At AWS re:Invent, Amazon Web Services, Inc. (AWS), an Amazon.com company, announced two new arrivals to complement its existing popular Amazon Elastic Container Service (Amazon ECS) and make it easier than ever to deploy, manage, and scale container workloads on AWS. Amazon Elastic Container Service for Kubernetes (Amazon EKS) brings Kubernetes to AWS as a fully managed service, making it easy for customers to run Kubernetes applications on AWS without the need to become experts in operating Kubernetes clusters. AWS also introduced a new capability called AWS Fargate that allows customers to launch and run containers without provisioning or managing servers or clusters.
“While we have over a hundred thousand active Amazon ECS clusters running on AWS and more customers running Kubernetes on AWS than on any other cloud, customers have also asked us to build a managed Kubernetes service like we have with Amazon ECS,” said Deepak Singh, GM of Containers and High Performance Computing Services, AWS. “Not only have we delivered on this request with Amazon EKS, but we’ve also made managed containers easier to use than ever before by launching AWS Fargate to allow developers to run containers at the task level rather than having to think about servers or clusters.”
Amazon EKS: The best way to run Kubernetes on AWS
Today, customers are running virtually every type of container orchestration and management service on AWS. In addition to Amazon ECS, Kubernetes has also become very popular with AWS customers. A recent survey from the Cloud Native Computing Foundation found that 63 percent of Kubernetes clusters running in the cloud are on AWS, more than any other cloud platform. Before today, operating Kubernetes with high availability required specialized expertise and a great deal of manual work. Customers needed to install and operate Kubernetes masters (which manage a customer’s clusters of servers) across multiple Availability Zones (AZs), replace unhealthy masters, and put measures in place to ensure that updates do not cause application downtime. Amazon EKS removes this complexity, making it easy for customers to run highly available Kubernetes environments. Amazon EKS is the first cloud service to deliver a highly available architecture that automatically distributes Kubernetes masters across multiple AZs to eliminate a single point of failure. This makes it easy for customers to deploy their applications in a highly available fashion. Applications running on Amazon EKS are resilient to the loss of a single master, or even an entire AZ. Amazon EKS automatically detects and replaces unhealthy masters, and it can automatically patch and perform version upgrades for masters.
With Amazon EKS, launching a Kubernetes cluster is as easy as a few clicks in the AWS Management Console. Amazon EKS handles the rest, automating much of the heavy lifting involved in managing, scaling, and upgrading Kubernetes clusters. Customers can run their existing Kubernetes applications on Amazon EKS without any code changes using existing Kubernetes tooling. In addition, customers get all the performance, scale, reliability, and availability of AWS, plus integrations with AWS networking and security services, including Application Load Balancer, AWS Identity and Access Management (AWS IAM), AWS PrivateLink, and AWS CloudTrail.
AWS Fargate – run containers without managing servers or clusters
Container orchestration services like Amazon ECS and Amazon EKS remove much of the heavy lifting involved in running containers at scale, but customers still need to provision and scale server instances and clusters, and patch the underlying Amazon EC2 instances. AWS Fargate makes running containers easier than ever by eliminating the need to manage clusters of servers. Customers no longer have to choose instance types, decide when to scale their clusters, or optimize cluster utilization. All customers have to do is define their applications as a ‘Task,’ which includes a list of containers, CPU and memory requirements, networking definitions, and AWS Identity and Access Management (AWS IAM) policies. Customers can launch thousands of Tasks in seconds and only pay for the resources in the Task—not for the infrastructure Tasks run on. AWS Fargate is available for Amazon ECS now and will be coming to Amazon EKS in 2018.
Realtor.com is a real estate listings website that allows potential home buyers and sellers to search real estate property records, houses, condos, and more online. “The transition to AWS Fargate was very smooth, simple, and fast,” said Jean Domiguez, Director of Cloud Services, Realtor.com. “This fundamentally changes how we deliver containers by removing the need to optimize container infrastructure and workload density. AWS Fargate is a game changer in container cluster management and delivery.”
Edmunds.com is an online resource for consumers to review car information for new and used automobiles. “AWS Fargate allows us to focus on developing and delivering features to our end-users while it manages the service management nuances,” said Ajit Zadgaonkar, Executive Director, Edmunds.com. “The abstraction at ‘Task’ level is a brilliant step towards making it easier to develop highly scalable microservices.”
AdRoll specializes in performance advertising marketing serving over 16K business-to-business clients worldwide. “We’ve migrated a number of applications into microservices, but there is still some overhead required to manage the clusters,” said Valentino Volonghi, CTO, AdRoll. “AWS Fargate gives even more teams the autonomy to quickly experiment with a variety of new services, without depending on an operations team. Basically, AWS Fargate lowers our operational cost to gain even more agility with containers.”
Expedia is a global travel company that operates several international online travel brands. “We’re excited to explore how AWS Fargate could help reduce the current operational overhead involved in managing Amazon ECS clusters,” said Matt Callahan, Engineering Manager for Cloud Automation, Expedia. “This includes monitoring and patching Amazon EC2 instances, tweaking cluster auto-scaling, and right-sizing instances. During the AWS Fargate customer beta testing we were impressed with the simplicity of creating a new serverless Amazon ECS cluster without the need for a complex AMI creation pipeline.”
