commitment

Spent some time today deconstructing the GFS paper from 2003. 

Reviewed the challenges around tensions between consistency and performance in distributed storage systems.

It continues to surprise me how basic very large systems are when broken down into their components.

Today I watched the lecture on primary-backup replication. Reviewed the pros and cons of using a direct state transfer compared to a replicated state machine and discussed various failure modes. Went through a detailed example of VMware's Fault Tolerance (FT) mechanism, which employs deterministic replay to synchronize a primary virtual machine with its backup.

Building /bright-moments - minting onchain art IRL | /purple #15 | FID 129 #magikarp

Rome, Metropolitan City of Rome Capital, Italy

Distributed systems are fun. To me, anyway. Some people find them too scary. It’s one of the reasons I left my last job. They were too frightened to move away from an ancient monolith because “distributed systems are hard”. But they had one anyway! Because it was a Citadel at that point and all the problems of distributed systems were *already there*. They didn’t trust their engineers. Just too conservative and cowardly for my liking. I like leading teams to tackle hard problems, not just do the easiest thing and spend decades dealing with the technical debt. (Oops! Sorry for the rant, I guess I have feelings about this. 😀)