Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.603760
Title: Rate limiting in an event-driven BGP speaker
Author: Harris, E. D.
Awarding Body: University of Cambridge
Current Institution: University of Cambridge
Date of Award: 2009
Availability of Full Text:
Full text unavailable from EThOS. Please contact the current institution’s library for further details.
Abstract:
The Border Gateway Protocol (BGP) is the Internet’s interdomain routing system  BGP is a global, distributed, peer-to-peer system consisting of tens of thousands of routers in the different networks which make up the Internet. Each BGP-speaking router, or speaker, is continually exchanging information with its neighbours about changes to the destinations to which it can send packets and the routes it uses to get there. However BGP is also just another application, competing for processor time and resources with all the other applications running on the same router. The way a router’s BGP software is implemented can have an enormous impact not only on that router’s individual performance, but on the performance of the Internet as a whole. Rate limiting is an important optimisation which can improve BGP’s network-wide performance and reduce its demands on the router running it. Nevertheless XORP, an interesting new software router aimed at both researchers and commercial users, does not support rate limiting. This is at least partly because XORP uses an innovative event-driven, pipelined route processing model which does not fit well with the traditional, timer-based way of implementing BGP with rate limiting. This dissertation argues that a serious, modern BGP speaker such as XORP must support rate limiting. It presents two different implementations of rate limiting in XORP’s route processing pipeline, representing different trade-offs between features and performance on the one hand and implementation impact t on the other. The first implementation requires no changes to the pipeline architecture but sacrifices some of its run-time capabilities. The second implementation requires fundamental changes throughout the pipeline but retains much more of its power. In the best case, experiments with a single XORP router show that both rate-limiting implementations significantly reduce BGP processing requirements; in the worst case, they equal or out-perform the standard XORP BGP speaker in all but one test. Further experiments, using several router instances in a virtualised testbed network, show that rate limiting also improves XORP’s network-wide behaviour.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.603760  DOI: Not available
Share: