Recap of AWS re:Invent – “re:Inforces” the Need to Test, Test, Test

By Todd Law On December 6, 2017
Cloud, networking, network testing

It was very exciting to attend last week’s AWS re:Invent conference and trade show in Las Vegas. This event has exploded in magnitude from a modest 6,000 attendees in its earliest incarnations, to an estimated attendance of around 40,000 people in 2017. Besides the trade show for AWS partners and sponsors (including Spirent), there were also over 1,000 technical sessions on diverse aspects of the AWS ecosystem. It’s hard to argue with these kinds of numbers, not to mention AWS’s growing market share in an exploding cloud marketplace.

For me, the highlights of the show were the keynote speeches, where a dizzying array of new features and services were announced. While voice, server-less architectures, and machine learning were certainly the dominant themes, it was also re-assuring to hear Amazon’s strong commitment to test.

test test test backdrop for speaker at AWS re-invent

AWS’s CTO, Werner Vogels, echoed this sentiment when he repeated the mantra of “Test, Test, Test” during his keynote address. This was not some off-the-cuff comment thrown in for no reason—Vogels had a definite reason for everything he said during his keynote.  In this case, Vogels’ comments are certainly motivated by the dramatic changes that cloud-based architectures bring to the world of software development. To elaborate, Vogels also made the telling statement, that, “All the code you’ll ever write will just be business logic”.

“All the code you’ll ever write will just be business logic.”
-- Werner Vogels, CTO at AWS

Think about that. The AWS ecosystem is now comprised of over 100 services.  Applications developed on AWS will ultimately be reduced to a set of business objects, attributes, relationships, and associated methods—everything else gets farmed out to services.  And AWS’s list of services is quite comprehensive, covering everything from baseline things like databases, compute, storage, etc., all the way up to exotic pieces like machine learning. And behind each of those services, is ultimately a server or multiple servers. Even so-called “server-less” offerings, such as AWS’s Lambda, actually use servers; however, server management and capacity planning tools are simply hidden from the developer or operator. And guess what, all those servers are going to have to communicate with the business logic.

While this is all really cool and super powerful for developers, it also raises some basic questions. What will the latency be across the network between all those servers? If I am using traditional servers, then how many servers do I need? If I am going server-less, will AWS’s cloud infrastructure rapidly scale as my applications’s needs also scales? These are the kinds of questions that Spirent can help you answer with products like Spirent CloudStress and MethodologyCenter, which uniquely, can simulate workloads and measure resource utilization in the cloud.

comments powered by Disqus
× Spirent.com uses cookies to enhance and streamline your experience. By continuing to browse our site, you are agreeing to the use of cookies.