An asymptotically optimal policy and state-space collapse for the multi-class shared queue

TL;DR

This paper develops an asymptotically optimal control policy for multi-class shared queues in heavy traffic, utilizing Harrison-Taksar free boundary solutions, with implications for cloud computing scheduling.

Contribution

It introduces a novel queuelength control policy derived from Harrison-Taksar problem solutions, differing from traditional priority rules.

Findings

The policy is asymptotically optimal in heavy traffic.

The queuelength control follows a triangular form.

Probabilistic methods are effective for cloud scheduling problems.

Abstract

We consider a multi-class G/G/1 queue with a finite shared buffer. There is task admission and server scheduling control which aims to minimize the cost which consists of holding and rejection components. We construct a policy that is asymptotically optimal in the heavy traffic limit. The policy stems from solution to Harrison-Taksar (HT) free boundary problem and is expressed by a single free boundary point. We show that the HT problem solution translated into the queuelength processes follows a specific {\it triangular} form. This form implies the queuelength control policy which is different from the known $c\mu$ priority rule and has a novel structure. We exemplify that the probabilistic methods we exploit can be successfully applied to solving scheduling and admission problems in cloud computing.