AWS Re:Invent 2017 – Serverless Containers, Managed Kubernetes, Bare Metal, Machine Learning, and IoT

Dec 1, 2017 9:56:21 AM

Breaking the Triangle of Cost, Quality, and Speed

This year’s AWS Reinvent delivered major announcements in DevOps, machine learning and IoT. All of the announced capability aim to eliminate infrastructure as the bottleneck for enterprises to become ‘digital attackers’. Observing the nearly 50.000 developers, architects, and software operators that came to Reinvent showed us a significant degree of genuine excitement about Amazon helping enterprises release new software faster, at a higher quality and lower cost.

AWS Fargate - ‘Serverless Containers'

AWS Fargate offers ‘serverless containers.’ With Fargate customers no longer need to worry about right sizing, scaling, load-balancing, monitoring, and patching EC2 instances. Instead of selecting a T-shirt size for a VM instance, users define the needed vCPU and vRAM for their container workload and only pay for the time these resources are used. Containers can still access any AWS service via Amazon’s Elastic Network Interfaces.

Containers now run directly on Fargate resources, instead of on EC2 instances

Availability: Available today in US East (Northern Virginia) region

Competitive Significance:  Based on EMA research, 40% - 50% of EC2 instances are either underutilized or entirely unused. With per second billing and a drastic reduction of operating requirements, Fargate could eliminate a signifiant share of this waste, enabling enterprise customers to pour the saved funds into developing new software.

Fargate goes significantly beyond Microsoft’s and Google’s offerings in that it now fully abstracts microservices from their underlying infrastructure. This follows the Lambda-principle where users only pay for the resources (RAM and CPU) they use on a per second basis, without the need to forecast resource requirements. This offers the same near-infinite scalability as Lambda, but with the flexibiliy of bringing the entire runtime within a container.

Managed Kubernetes - Amazon Elastic Container Service for Kubernetes

With Amazon Elastic Container Service for Kubernetes (EKS), Amazon finally delivered a managed Kubernetes offering that eliminates the requirement to stand up and operate Kubernetes clusters. Load balancing, high availability, scalability, and upgrading are key Kubernetes pain points that require staff skills that are often not present in an enterprise, preventing development teams from fully taking advantage of containers. EKS leverages upstream Kubernetes to ensure that any Kubernetes app and plugin will work without modification.

AKS takes care of providing the Kubernetes infrastructure in a highly available manner

Availability: You can sign up for a tech preview here: https://pages.awscloud.com/amazon-eks-preview.html

Competitive Significance: EKS was much expected and is equivalent to what Microsoft and Google are already offering. However, connecting EKS with the rest of the AWS empire brings possibilities that the competition will have a hard time matching.

Amazon EC2 Bare Metal Instances

With EC2 Bare Metal Instances Amazon is filling a gap that prevented the move of applications that require direct access to local hardware capabilities, such as graphics cards or processor command sets, are subject to rigid licensing or support requirements, or simply perform better on bare metal. Customers could even install their own hypervisor platform.

Availability: Public Preview (i3.metal instance) today. Rollout during 2018.

Competitive Significance: Customers can now run their bare metal apps in the Amazon Cloud and connect them to any of the hundreds AWS services. This brings Amazon ahead of both of its key competitors, Google and Microsoft.

Improved Deployment Options for AWS Lambda - AWS Code Deploy now Supports Lambda

- AWS Code Deploy can now push out Lambda functions: This makes blue/green testing on Lambda possible and therefore enables a gradual rollout of new functions. With Code Deploy, users can also automate the build, test, deploy, and release process for serverless functions.
- Serverless repository: [https://aws.amazon.com/serverless/serverlessrepo/]
- 3008MB memory maximum replacing the old 1,536MB max for serverless functions
- Concurrency limit to throttle cost

Availability: Now

Competitive Significance: Making Lambda part of the DevOps pipeline is critical. Of course, most vendors use third party tools for DevOps pipeline automation, but by integrating Lambda with CodeDeploy, Amazon at least points the way into the right direction.

We will look at improvements in machine learning and IoT in a separate post.

Torsten Volk

Written by Torsten Volk

With over 15 years of enterprise IT experience, including a two-and-a-half-year stint leading ASG Technologies' cloud business unit, Torsten returns to EMA to help end users and vendors leverage the opportunities presented by today's hybrid cloud and software-defined infrastructure environments in combination with advanced machine learning. Torsten specializes in topics that lead the way from hybrid cloud and the software-defined data center (SDDC) toward a business-defined concept of enterprise IT. Torsten spearheads research projects on hybrid cloud and machine learning combined with an application- and service-centric approach to hyperconverged infrastructure, capacity planning, intelligent workload placement, public cloud, open source frameworks, containers and hyperscale computing.

  • There are no suggestions because the search field is empty.

Lists by Topic

see all

Posts by Topic

see all

Recent Posts