Abstract
Data-over-cable service interface specifications (DOCSIS),
the de facto standard in the cable
industry, defines a scheduling service called real-time polling service (rtPS)
to provision quality of service (QoS) transmission of real-time variable bit
rate (VBR) videos. However, the rtPS service intrinsically has high latency,
which makes it not applicable to real-time traffic transport. In this paper,
we present a novel traffic scheduling algorithm for hybrid fiber coax (HFC)
networks based on DOCSIS that aims to provide QoS for real-time VBR video
transmissions. The novel characteristics of this algorithm, as compared to
those described in published literatures, include 1) it predicts the bandwidth
requirements for future traffic using a novel traffic predictor designed to
provide simple yet accurate online prediction; and 2) it takes the attributes
of physical (PHY) layer, media access control (MAC) layer and application
layer into consideration. In addition, the proposed traffic scheduling algorithm
is completely compatible with the DOCSIS specification and does not require
any protocol changes. We analyze the performance of the proposed traffic predictor
and traffic scheduling algorithm using real-life MPEG video traces. Simulation
results indicate that 1) the proposed traffic predictor significantly outperforms
previously published techniques with respect to the prediction error and 2)
Compared with several existing scheduling algorithms, the proposed traffic
scheduling algorithm surpasses other mechanisms in terms of channel utilization,
buffer usage, packet delay, and packet loss rate.
© 2009 IEEE
PDF Article
More Like This
Cited By
You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
Contact your librarian or system administrator
or
Login to access Optica Member Subscription