Why Serverless Computing? (Part 2 of 5)


Part 2 of a 5 Part Series. See the other videos here: All About Serverless Computing

So the reason we might want to implement serverless computing isn't so much because we think there is a great improvement in performance. There's not even really a significant improvement. If you just look at the costs of running functions versus owning a cloud instance and running a server running your application. If the functions are running all the time it's probably cheaper in from an infrastructure perspective or cloud cost perspective to buy the instance and run your VMs. The real thing that comes down to that I've been hearing is about the total cost. Once you go serverless your IT shop can shrink down to really kind of a minimum staff that focuses on business value because they're no longer doing operations. You've outsourced all operations costs so even running Kubernetes has an operational cost. So people think containers are the end all be all and like no there's people that are saying out there look I don't even want to run Kubernetes and be woken up in the middle of night because my Kubernetes servers hung up or something. If I get rid of all of that operations costs I can really run a cheap operation. And likewise development costs become extremely agile and you can focus very quickly on adding little functions around the edges and re-accumulating those functions to make new applications much faster than you can re-architect a version 2 of a somewhat larger client server application or...in heaven forbid a monolithic application where you have to really lift and shift it. So, it's a brave new world and the people who are doing it saying look it really is a cost savings at the end of the day.