Use this URL to cite or link to this record in EThOS:
Title: Operational support systems for satellite communications
Author: Fitch, Michael Robert
ISNI:       0000 0001 3467 7499
Awarding Body: Open University
Current Institution: Open University
Date of Award: 2005
Availability of Full Text:
Access from EThOS:
Access from Institution:
The role of satellite communications is changing from providing bandwidth linking network operators interconnections towards providing IP enabled communications to end users. This migration from few high-value routes towards many low-value routes means that integration and automation of processes with terrestrial networks becomes critical in driving down unit costs. Integration and automation is necessary on all planes: user, control and management. In satellite communications, management aspects, underpinned by Operational Support Systems (OSS) have received the least research attention, making this a valuable topic for study. In most areas, OSS for satellite systems are similar to other domains. However there are some notable areas of difference which have been the focus of this research. The eTOM business framework, developed by the TMF, has been used to highlight aspects of OSS unique to satellite. Since satellite capacity represents the highest operational cost of a satellite route, effective management while minimising the overhead traffic is critical. The transmission of IP packets is assumed and the real-time measurement of QoS parameters such as packet delay and loss emerged as the most important differences. A number of approaches to QoS measurement are feasible, however the use of trace packets is most promising especially for high network loads. An experiment compares the results from simulations, mathematical models and from a test network, using Poisson and self-similar traffic flows. The relationship between measurement accuracy and trace packet intensity is explored and the measurement response time to steps in traffic load is estimated. It is discovered that measurement accuracy improves as the queue load increases, in contrast to alternative approaches such as sampling of user packets. The response time to steps depends upon the degree of self-similarity and is generally longer than the times recommended by standards. A pragmatic approach to management of different modes is proposed where the measurement method is changed depending on the load.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral