Disorder & Tolerance in Distributed Systems at Scale
summary
Argues that distributed systems can borrow biological principles — ensemble forecasting for multiple failure modes, entropy as a measurable snapshot, and replication, regeneration, rebalance as native resiliency tactics — and shows a concrete case: Yellowstone loses its wolves, elk climb to ~19,000, reintroducing 14 then 17 more wolves grows the population to 132 by 2001 and restores forests and rivers. Suggests embedding self-organization and diversity rather than a single external orchestrator to improve resilience at scale.
AI-written summary · about
chapters (5)
- 0:00Introduction and motivation
- 10:00Complexity theory in systems
- 20:00Biological systems and resilience
- 30:00Yellowstone case study and lessons
- 38:00Innovation and future directions
related talks
33mWhy Are Distributed Systems So Hard?
41mOnce Upon a Vulnerability: Building Resilience in a Chaotic World
25mGrowing Distributed Systems
44mDistributed systems and scalability rules
22mWhy Are Distributed Systems so Hard?
27mDistributed Systems, Like It or Not
28mWhy Are Distributed Systems So Hard?
51mPatterns for building resilient software systems