Protego: Overload Control for Applications with Unpredictable Lock Contention
summary
Protego implements an Active Synchronization Queue Management (ASQM) that aborts requests whose remaining lock-wait budget cannot meet SLOs, plus a credit-based marginal-throughput admission control that nudges load based on measured efficiency (responses/requests) instead of CPU or queue depth. It runs on real apps (Lucene, Memcached), drops roughly 20% of requests in stressed setups, and reports up to 3.3x more goodput and 12.2x lower 99th‑percentile latency versus prior overload controllers.
AI-written summary · about
chapters (8)
- 0:00Introduction to Protego
- 0:36Challenges in data centers
- 2:44Impact of lock contention
- 6:36Protego architecture
- 8:54Performance analysis
- 10:57Efficiency-based control
- 12:39Evaluation and results
- 16:16Conclusion
related talks
25mAvoiding Overload in Distributed Systems
20mBetter Resilience Through Self Tuning
40mEverything Will Flow
25mResQ: Enabling SLOs in Network Function Virtualization
45mPerformance-by-Design: Embedding Intelligent Scaling and Guardrails Into Platform Engineering
32mTaking Care of Your Control Plane With API Priority and Fairness and Resource Quotas
52mBulletproof Java Enterprise Applications for The Hard Production Life
32mHow does your application behaves when everything goes wrong?