ANALYSIS: Amazon Web Services wants enterprises to get ready to migrate to a world of micro services running in a “serverless” cloud universe
Amazon Web Services is showing it has not run out of ways to improve or expand the AWS cloud, especially for application developers.
At the company’s New York City Summit here Aug. 11 at a packed Javits Center, AWS CTO Werner Vogels discussed several developer initiatives based on its contention that the cloud’s economic model makes more sense than ever for enterprises.
Vogels presented updates on micro services initiatives and the Lambda serverless computing architecture, both of which he sees as the future of application development on AWS, along with containers.
Micro services are big applications broken down into fundamental parts so they can run in containers on the Lambda architecture. They are serverless apps in the sense that they run in memory on the cloud platform.
He said AWS is adding a Usage Plan option to its API Gateway service that gives developers more control over their APIs, and helps them develop Lambda serverless apps. And developers need the help, company officials explained.
“Serverless apps could end up as complicated as the monolithic applications they are replacing,” said Dr. Tim Wagner, general manager for AWS’ serverless initiative. In a session, he discussed the developer preview of Project Chalice, a Python-based serverless “microframework,” and another initiative, Project Flourish, which provides application templates for serverless code.
Both are ways to “bring order to the chaos,” Wagner said. The idea is to be able to deploy app code like containers, mix and match the app components, and be able to run services independently of each other as needed, he said.
As it has demonstrated over the years, AWS is keeping ahead of customers in offering new services. Most enterprises are nowhere near ready to adopt micro services or deploy serverless code, but cloud-native AWS customers like Lyft and Airtime are benefiting from the ability to create more flexible applications more quickly and with more scalability.
Airtime is a startup that has created a real-time social video and sharing service on AWS. The company had a lot of problems initially building the social video site, explained Abby Fuller, a DevOps engineer for Airtime. “It was monolithic, hard to change, and single issues could cause systemwide failures,” she said. “We had to rebuild the whole thing.”
Vogel asserted that in the serverless world, we have gone from managing “pets” (individual servers) to “cattle” (commodity servers in the cloud) and now the “herd” (serverless code). In a serverless world, hardware becomes invisible. “There’s only the application,” he said.
AWS also knows businesses need help getting to the cloud in a more traditional sense, so they are also pushing services that go beyond building new apps in AWS to helping enterprises move existing applications to the cloud or enabling them to work with the cloud.
To support these efforts, the cloud company is offering the AWS data migration appliance, Snowball. Other migration support components include the new AWS Snowball Job Management API and also an S3 (Simple Storage Service) Adapter that lets IT administrators save data to a Snowball box as if it were an S3 endpoint.
Vogels also announced the new Application Load Balancer, which performs for applications what the Elastic Load Balancer does for compute nodes. Application Load Balancer enables control over how to send traffic to individual components, content-based routing, HTTP/2 support and detailed logging, Vogels said.
Other announcements included new IPv6 support, with S3 data stores now accessible via IPv6 addresses. Also, in security, the Key Management Service now supports bring-your-own encryption keys.
It’s all in the name of getting enterprises looking at the cloud not as the future, but the immediate future, and helping them get there as soon as possible.
“It’s all about a journey,” Vogel said, and everybody is going to have to take it.
Originally published on eWeek.