Push Cache

I know, when I start talking high-finance, everyone loses their fucking minds.

As 5G and edge-compute becomes prevalent, it would be nice if there were technologies that made it easy to push content to various cache-levels.  The biggest difference people see in application optimization is moving to a push-driven data flow.  Whenever data updates, it should propagate to fast, distributed caches to minimize round-trips to databases and expensive functions.  This design pattern, with a strong network of edge locations, is what allows us to deploy lightning fast real-time applications in the cloud.

This async-first design mindset is all the rage.  It’s got the programming languages and the silicon too, now, and for good reason.  Changing the impedance of your systems to have data flow out when it updates, rather than consumers polling for new data, is an excellent way to optimize any application.  If you play around with it you can find just the right timings for cache expiration at different points of presence.  Hopefully we will see automated tooling for tuning edge cache soon.  Edge compute can work a lot like a mesh network, but it’s difficult to stand these systems up unless you have quality orchestration tools.