Yash

    Research

    Measured, with the data released.

    Papers and studies on how backend systems behave under load. Every number comes from experiments I ran, and the code and raw data are public so anyone can rerun them.

    Papers

    1. Preprint, Zenodo15 pages

      Overshoot, Retries and Clocks: An Empirical Study of Redis-Based Rate Limiting Under Concurrency

      Yash Suthar · doi:10.5281/zenodo.23467402

      Check-then-act (GET, compare, INCR) let through 29% more than its limit on average with 64 workers, and 37% in the worst trial. Under network delay that overshoot follows a lockstep model, c·ceil(L/c) − L, to within about one request.

      Abstract and citationPDFCode and data (.zip)Short version

    Studies and data

    1. Atomic rate limiting with Redis and Lua: cost, memory, blocking and cluster behaviour

      Seven experiments on Redis 7.0: the CPU cost of a scripted rate limiting decision, sliding-log memory per client, how long large sorted sets block Redis, EVAL vs EVALSHA vs FCALL across restarts, the timestamp-member bug in sliding logs, Redis Cluster key rules, and replication of TIME. Includes the Go harness and raw data.

      Read the resultsCode and data (.zip)

    Also on ORCID. More notes from building these systems are in Writing.