Web Object Size satisfying Mean Waiting Time In Multiple Access Environment

TL;DR

This paper derives the mean web object size that satisfies specific waiting time constraints in multiple access environments using queueing models, aiding in web service design and performance optimization.

Contribution

It introduces a novel approach to determine web object size based on queueing theory models like M/G/1, M/D/1 with vacations, and M/H2/1, considering different scheduling and access environments.

Findings

Mean web object size increases with link utilization.

Size converges to a lower bound beyond a threshold number of objects.

Results assist in economic web service design and maintenance.

Abstract

This paper addresses web object size which is one of important performance measures and affects to service time in multiple access environment. Since packets arrive according to Poission distribution and web service time has arbitrary distribution, M/G/1 model can be used to describe the behavior of the web server system. In the time division multiplexing (TDM), we can use M/D/1 with vacations model, because service time is constant and server may have a vacation. We derive the mean web object size satisfying the constraint such that mean waiting time by round-robin scheduling in multiple access environment is equal to the mean queueing delay of M/D/1 with vacations model in TDM and M/H2/1 model, respectively. Performance evaluation shows that the mean web object size increases as the link utilization increases at the given maximum segment size (MSS), but converges on the lower bound when the number of embedded objects included in a web page is beyond the threshold. Our results can be applied to the economic design and maintenance of web service.