QCluster: Clustering Packets for Flow Scheduling

TL;DR

QCluster is a fast, switch-implemented clustering framework that adapts flow scheduling in data centers, significantly reducing flow completion times by efficiently grouping packets with similar properties.

Contribution

It introduces QCluster, the fastest known clustering algorithm for flow scheduling, capable of real-time packet clustering in commodity switches with limited queues.

Findings

Reduces average flow completion time for short flows by up to 56.6%.

Decreases overall average flow completion time by 21.7%.

Operates at 3.2 Tbps in Tofino switches, demonstrating high-speed clustering.

Abstract

Flow scheduling is crucial in data centers, as it directly influences user experience of applications. According to different assumptions and design goals, there are four typical flow scheduling problems/solutions: SRPT, LAS, Fair Queueing, and Deadline-Aware scheduling. When implementing these solutions in commodity switches with limited number of queues, they need to set static parameters by measuring traffic in advance, while optimal parameters vary across time and space. This paper proposes a generic framework, namely QCluster, to adapt all scheduling problems for limited number of queues. The key idea of QCluster is to cluster packets with similar weights/properties into the same queue. QCluster is implemented in Tofino switches, and can cluster packets at a speed of 3.2 Tbps. To the best of our knowledge, QCluster is the fastest clustering algorithm. Experimental results in testbed with programmable switches and ns-2 show that QCluster reduces the average flow completion time (FCT) for short flows up to 56.6%, and reduces the overall average FCT up to 21.7% over state-of-the-art. All the source code in ns-2 is available in Github without.