Can Microservices Drive a Renaissance in Workload-Aware Storage Management?
Talkat 12th USENIX Workshop on Hot Topics in Storage and File Systems (HotStorage '20) · 2020-07-13..07-14·
summary
Paper evaluates microservices using docker and kubernetes with a per-volume io cache and shows a concrete auto-tuning prototype: a collector and recommender pipeline that gathers per-persistent-volume traces and writes recommendations into extended artifacts in the repo. Experimental comparison uses arc and lru on two workloads (virtual machines and a production customer database) and finds isolated per-persistent-volume cache beats a shared cache by a large margin, and that unequal cache split ratios can further improve hit rate.
AI-written summary · about
chapters (7)
- 0:00Start
- 1:45Research
- 2:05Motivation
- 3:01Workload Stability
- 3:51Shared vs Isolated Cache
- 4:28Cache Size Allocation
- 4:55Initial Design
related talks
48mChallenges in Implementing MicroServices
37mKnee Deep in Microservices
20mMicroservices in the future
26mEnhance System Resilience and Efficiency with Microservices
45mMicroservices: A retrospective
23mMicroservices and containers
47mIt is not just Microservices
49mDeveloping microservices applications on the Kubernetes platform
