Class Central is learner-supported. When you buy through links on our site, we may earn an affiliate commission.

YouTube

On Modular Learning of Distributed Systems for Predicting End-to-End Latency

USENIX via YouTube

Overview

Explore a cutting-edge approach to modeling end-to-end latency in distributed systems through this 19-minute conference talk from NSDI '23. Discover Fluxion, a novel framework that employs modularized learning to characterize system performance more efficiently and accurately than traditional monolithic models. Learn how this innovative method significantly reduces adaptation costs in dynamic cloud deployments, allowing for faster model updates and improved system performance tuning. Gain insights into the implementation of learning assignments, a new abstraction that enables modeling of individual sub-components with a consistent interface. Examine the framework's effectiveness through case studies involving complex systems with up to 142 microservices, demonstrating substantial improvements in prediction accuracy and system optimization.

Syllabus

NSDI '23 - On Modular Learning of Distributed Systems for Predicting End-to-End Latency

Taught by

USENIX

Reviews

Start your review of On Modular Learning of Distributed Systems for Predicting End-to-End Latency

Never Stop Learning.

Get personalized course recommendations, track subjects and courses with reminders, and more.

Someone learning on their laptop while sitting on the floor.