Murdoch University Research Repository

Welcome to the Murdoch University Research Repository

The Murdoch University Research Repository is an open access digital collection of research
created by Murdoch University staff, researchers and postgraduate students.

Learn more

Incorporating deadline scheduling into DCCP

Wilson, Daniel (2013) Incorporating deadline scheduling into DCCP. PhD thesis, Murdoch University.

PDF - Front Pages
Download (149kB)
PDF - Whole Thesis
Download (2MB)


In current TCP, UDP and DCCP networks, the finite nature of real time data is not taken into consideration when scheduling of packets occurs on intermediate devices. To overcome this, the research presented in this thesis incorporates deadline scheduling into DCCP. This allows packet age information to be transferred by the DCCP protocol which provides intermediate devices with the ability to make more informed transportation decisions for real time data. By embedding two variables, in the form of options into DCCP-Data packets, a number of new prioritization and scheduling mechanisms that utilize packet age information are made possible. This thesis will show how deadline scheduling is incorporated into DCCP in a stable, backward compatible and DCCP standard complaint manner.

Once deadline scheduling is incorporated into DCCP, the focus of the thesis then shifts towards mechanisms that can be implemented on intermediate devices that make use of the packet life information. To begin, five unique packet discard mechanisms that purge stale packets from the network are presented. The purpose of these mechanisms is to remove stale packets from the network using intermediate devices in the network in order to free network resources for non stale packets utilizing the same infrastructure. Experimentation carried out to investigate the efficiency and benefits to DCCP performance offered by each of these five mechanisms is shown. The experimentation also explores fairness amongst competing flows when the mechanisms are activated.

Following this, a novel probabilistic scheduling (PBS) mechanism is introduced that predicts the probability a packet has of arriving at its intended destination net- work within its useful lifespan. Once this probability is calculated scheduling decisions are then made based on this value in order to offer optimized delivery to real time data. In order to calculate this probability, the PBS mechanism utilizes metrics from the Cisco EIGRP routing protocol. Experiments carried out using the PBS mechanism demonstrate that the PBS mechanism improves DCCP performance in networks where high levels of stale packets occur.

Overall, this research presents a new approach to the transportation of real time data in DCCP networks and aims to improve DCCP adoption through the improved performance capability added to the protocol by this research.

Item Type: Thesis (PhD)
Murdoch Affiliation(s): School of Engineering and Information Technology
Supervisor(s): Dixon, Michael and Koziniec, Terry
Item Control Page Item Control Page


Downloads per month over past year