Event: Bengaluru Systems Meetup #4

Had an incredible time presenting at this meetup! Interestingly, this topic was actually one of the first I presented almost 5 years ago during the SRE Learning Sessions at Zomato.

Distributed Systems are born out of the need to support scale beyond a single node. Consequently, balancing the load among all the nodes in a distributed system should be a core functional attribute of that system.

But in reality very few operators monitor load variance in the distributed systems they manage. (unless of-course there is an incident where max utilisation of a single node starts a cluster wide incident or when you start optimising the cost of the system).

I presented various types of distributed systems and how load variance creeps into it easily and how some systems design itself incorporates “Anti Load Variance” to a large degree.

Slide Deck