Efficient Replication of Queued Tasks for Latency Reduction in Cloud Systems

TL;DR

This paper analyzes how different redundancy strategies in cloud systems impact latency and cost, revealing that the effectiveness depends on the log-concavity or convexity of service time distributions.

Contribution

It provides a theoretical framework for optimizing redundancy strategies based on the service time distribution's properties, balancing latency and cost.

Findings

Log-concavity of service time distribution influences redundancy effectiveness.

Maximum redundancy reduces latency and cost for log-convex distributions.

Fewer replicas and early cancellation are optimal for log-concave distributions.

Abstract

In cloud computing systems, assigning a job to multiple servers and waiting for the earliest copy to finish is an effective method to combat the variability in response time of individual servers. Although adding redundant replicas always reduces service time, the total computing time spent per job may be higher, thus increasing waiting time in queue. The total time spent per job is also proportional to the cost of computing resources. We analyze how different redundancy strategies, for eg. number of replicas, and the time when they are issued and canceled, affect the latency and computing cost. We get the insight that the log-concavity of the service time distribution is a key factor in determining whether adding redundancy reduces latency and cost. If the service distribution is log-convex, then adding maximum redundancy reduces both latency and cost. And if it is log-concave, then having fewer replicas and canceling the redundant requests early is more effective.