As recently reported in the Wall Street Journal (10-27-16), expect the next wave of computing that will be widely adopted later this decade as the cloud grows more ubiquitous to be “serverless” computing.
The word is so new, spell-checkers don’t even know about it yet. As GE’s CTO, Chris Drumgoole commented, it’s what the “cool kids” are thinking about.
Serverless computing is the logical evolution of physical and virtual servers, where the servers used to run applications become invisible to the developers building applications. GE’s Drumgoole expects the technology to really take off next decade.
Serverless computing “allows developers to focus only on writing code without having to manage the servers” – essentially make the process “serverless” to the customer.
As the Journal’s Sara Castellanos describes it: “The provider runs the customer’s application on its own servers or inside “containers,” where they are broken into small pieces and placed into software shells, allowing the pieces to be distributed to any sort of device, in a digitally orchestrated manner, and at lower cost.”
Speculation is that customer IT departments might then “pay the provider every time their code is triggered, instead of doling out cash upfront for machines or virtual servers that they may not need.” This is not unlike the way Amazon (via AWS) and Microsoft (via Azure) offer up their own cloud services to users today.
It’s the logical evolution begun a few years ago as physical servers began giving way to virtual ones.
Pioneers in the area today already include IBM, Alphabet (Google) and Amazon Web Services. The AWS offering, called Lambda, allows software developers to write applications and upload their code to its service, where Lambda manages the code and runs it on AWS servers. It executes that code at scale and bills for every millisecond the code is triggered, thus eliminating the need for customers to buy or maintain any physical servers.
It’s all part of the continuing shift to the cloud. The Journal concludes by pointing out that “it is particularly useful when running applications for internet-connected devices such as Amazon’s Echo, because the apps require massive amounts of requests of short duration.”
Welcome to the future.