Loading…
A Scheduling Algorithm for Bursty Traffic: Controlling of Service Rate and Burst
Quality of service (QoS) provisioning in communication networks highly depends on packet scheduling algorithms. In a well-known group of scheduling algorithms, i.e., Rate proportional servers (RPS), service rate of each session is used to isolate service disciplines. Some type of services, e.g., vid...
Saved in:
Published in: | Arabian Journal for Science and Engineering 2014-06, Vol.39 (6), p.4753-4764 |
---|---|
Main Authors: | , , |
Format: | Article |
Language: | English |
Subjects: | |
Citations: | Items that this one cites Items that cite this one |
Online Access: | Get full text |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Summary: | Quality of service (QoS) provisioning in communication networks highly depends on packet scheduling algorithms. In a well-known group of scheduling algorithms, i.e., Rate proportional servers (RPS), service rate of each session is used to isolate service disciplines. Some type of services, e.g., video streaming and interactive gaming not only have a bursty traffic but also need a minimum delay bound to obtain an acceptable QoS. However, unfortunately such applications may impose an unacceptable delay when the RPS scheduling algorithms are applied. In this paper, we propose a scheduling algorithm, which attempts to involve burstiness in addition to service rate, for isolating service disciplines in the scheduler. The proposed algorithm belongs to fluid flow paradigm and assigns a time variable weight that indicates the instantaneous service rate, to each session. The arrival constraint is assumed to be leaky bucket and our algorithm tries to provide a service discipline similar to the arrival constraint. Evaluation of the algorithm is carried out by calculating packet delay through a simulation where various kinds of traffic are scheduled by the proposed algorithm. The simulation results are compared with a well-known RPS scheduling algorithm and show that average, maximum and variance values of delay are more controllable in our algorithm by adjusting the parameters of each session. |
---|---|
ISSN: | 1319-8025 2191-4281 |
DOI: | 10.1007/s13369-014-1086-7 |