The production team has been looking at some optimizations this past cycle to improve our resource usage and reliability. Since we started backing the Samrong child chain and watcher with Parity instead of geth, we’ve seen better reliability and needed to do significantly less wrestling to keep things running. However, we were still seeing high resource usage and spent some time reducing the number of RPC calls we make to Parity to see if that would change Parity’s memory footprint. We continue to work to improve our metrics and monitoring so that we meet production service standards while operating in PoA. We believe this is critical as we onboard potential projects to use the plasma chain. We believe that availability is a required feature to projects that are serious about trying to integrate to the OMG Network.
This cycle, we’ll be putting our availability and reliability work to the test while we prepare the Plasma Dog Arena event that Hoard will be hosting this month. We’ll set up a test environment to conduct a stress test and see how much load we can throw at it. More details on this will be available in coming weeks.
We’ve also broken ground on the plasma abstract contracts implementation. If you recall from my How’s OmiseGOing? post, this contract architecture will allow us to perform opt-in feature upgrades without requiring all of our users to exit and re-deposit into a new chain, as we’re doing now with the Ari to Samrong hard network upgrade. This will provide a significantly improved user experience while enabling us to more easily add features in a trustless manner.