Winter is coming but so is a new instance of PYtalks! All the usual applies. I hope to see as many of you as possible.
Ivan Mesic: High Availability’ RPC implementation in the cloud
When one builds a ‘big’* system, at some point it starts to matter how many calls you can make and who can handle the load. One part of the solution is to split your system into ‘microservices’ so you can divide the work (all hail: divide and conquer).
At that point one gets hit with load and availability problems. One nice solution is to use a queueing technology to pass calls and responses around instead of client/server talking to each other. With RPC implemented over queue, one can bring up as many servers/clients as is needed! It looks like one solved all of his load/availability problems..
We will look at one simple (and stupid) implementation to work out the basics and along the way see how new Python 3.5 `async` and `await` keywords help.
* big meaning `distributed`, whatever that one means.