18 docs tagged with "Distributed systems"

Availability is arithmetic before it is architecture. The arithmetic is worth doing first, because

An estimate is not a prediction. It is a filter: a way to find out, in two minutes and before

CAP is the most cited and most misquoted result in distributed systems. The popular version —

Rule: a dying dependency needs to be isolated, not attacked with more requests. Breakers stop

Rule: agreement requires a majority, not a coordinator. Any scheme where a single node decides —

A consistency model is a contract about what a read is allowed to return. It is not a quality

Rule: failure is the default state, not the exception. Large systems are not designed for the

Rule: ACID stops at the database boundary, and "localhost is a great lie." Across services you

The point of knowing these is not recall for its own sake. It is being able to reject a design in

This consolidates two earlier drafts — MPC-Rust.mdx and mpc-cluster-ssd.md — which

* Network messages can be delayed, reordered, or lost.

Rule: "eventually consistent" is not a promise you can hand a user. Each read path has to

Design record for recurring billing — creating a subscription, taking the first payment,

Open-ended by construction: "design Twitter" has no correct answer, and the interviewer is not

Rule: a healthy service plus a healthy service does not equal a healthy system. The network

Rule: retry without idempotency is not error handling — it is a bug. And a retry that fires

VSR is a state machine replication protocol designed to keep a group of replicas consistent even when some fail. It

1. Document Control