InfiniCache: Exploiting Ephemeral Serverless Functions to Build a Cost-Effective Memory
summary
InfiniCache runs entirely on AWS Lambda, shards objects with Reed‑Solomon (k+m) erasure coding, and coordinates via a proxy and client library to stream chunks to specific lambda instances; periodic warm‑ups and peer relay backups reduce reclamation risk. Measured on IBM container‑registry traces with ~400 Lambdas, it reports 31–96× cost savings versus ElastiCache, ~95.4% data availability per one‑hour window, comparable latency for large (>10 MB) objects, and a ~13 ms invocation cost that hurts tiny‑object latency.
AI-written summary · about
chapters (7)
- 0:00Introduction and challenges
- 5:31Serverless cache design
- 10:24Data operation workflows
- 15:35Maximizing data availability
- 19:06Performance and cost evaluation
- 22:27Latency and conclusion
- 24:56Q&A session
related talks
06mCache Is the King! Create Cache in AWS Lambdas
58mServerless: Five Key Things you need to Know
42mServerless Systems That Won't Burn a Hole in Your Pocket
15mAlternative Ways To Think About Serverless
54mServerless beyond Functions
40mWorking with AWS Lambda without Breaking the Bank
35mEfficiency of Serverless Architectures with AWS Solutions- 42mServerless Systems That Won't Burn a Hole in Your Pocket