Ep 5: Datacenter Architectures and Cloud Microservices with Dr. Christina Delimitrou, Cornell University
Computer Architecture Podcast
00:00
The Performance Improvement of the Sage System Using Machine Learning
Sage relies on machine learning, but is entirely unsupervised. Skale doesn't need as much instrumentation at the kernel level, even at the application level. Counter factuals are hypothetical senarios of what would happen to the entend application if i were to twiak something in one of the existing micro services. The accuracy that it achieves is pretty similar to seer,. which was good, better than we expected to usually supervise learning works better.
Play episode from 15:23
Transcript
