| dc.description.abstract | Computer networks have become an integral part during capture, processing, storage and transmission of mission critical data.. This has resulted to a new demand of more elaborate and strict service levels that guarantees user satisfaction. Provisioning and measurement of this need is therefore the endeavor of every networking and telecommunication design effort. The need may vary considerably from one user to another resulting to the many definitions of quality of service. To be able to address these demands. all user requests have to be captured, policies specified inline with the requests so that strategies of provisioning can be fonnulated and subsequently implemented harmoniously.
This research project setout to determine if network perfonnance can be positively altered without increase of bandwidth. This was done by altering default features of some select network edge devices in respect to quality of service then measuring and eValuating such parameters as average Ethernet delay, email response time, VolP end to end delay against another network which is similar but implemented using default settings. This procedure was made possible by the use of OPNET simulation software to model and compare the performance of the two networks.
The construction of the first network is based on aggregate information gathered from some select production networks and is a representation of the status of our campus networks. The second network is the proposed model which is a duplicate of the first in terms of application services, number of users and traffic levels but configured to implement a policy based quality of service while taking advantage of inbuilt edge devices capabilities.
The results of the research have shown that adopting a policy based quality of service provisioning can improve some aspects of network performance without any investment in bandwidth expansion or infrastructuraI upgrade. These results have demonstrated that if such a procedure is implemented in a production network, it would temporarily postpone bandwidth expansion which would subsequently cut on operations cost and increase user productivity. | en_US |
