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

YouTube

Credence: Augmenting Datacenter Switch Buffer Sharing with ML Predictions

USENIX via YouTube

Overview

Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only!
Explore a groundbreaking research presentation on improving datacenter switch buffer sharing through machine learning predictions. Delve into CREDENCE, a novel drop-tail buffer sharing algorithm that leverages ML predictions to emulate push-out buffer performance. Learn how this approach can significantly enhance throughput and reduce flow completion times in datacenter switches with shrinking buffer sizes. Discover the potential of CREDENCE to achieve near-optimal performance comparable to the Longest Queue Drop (LQD) algorithm while gracefully handling prediction errors. Gain insights into the practical implementation of this technique using off-the-shelf machine learning methods compatible with current hardware. Understand the implications of this research for future developments in both systems and theory within the field of datacenter networking.

Syllabus

NSDI '24 - Credence: Augmenting Datacenter Switch Buffer Sharing with ML Predictions

Taught by

USENIX

Reviews

Start your review of Credence: Augmenting Datacenter Switch Buffer Sharing with ML Predictions

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.