Measurement based optimal multi-path routing
- 23 December 2004
- proceedings article
- Published by Institute of Electrical and Electronics Engineers (IEEE)
Abstract
We propose a new architecture fnr efficient network monitoring and measurements in a traditional IP netw,ork. This new architecture enables estahlishmrnt nf niultiple paths (tunnels) hetween source-destination pairs without having to modify the underlying routing prntoeol(s). Based nn the proposed architecture we propose a measurement-based multi-path rmuting algorithni derived from simultaneous perturbation stochastic appmrimution. The proposed algorithm does not assume that the gradient of analytical cost function is known to the algorithm, hut rather relies on noky estiniates from nimsurenimts. Using the analytical model presented in the paper we prove the canvergenee of the algorithm to the optimal solution. Simulation results are presented to denionstrate the advantages of the proposed algorithm under a variety of network scenarios. A comparative study with an existing nptimal routing algorithm, MATE, is also provided. a domain without requiring major changes in the infrastruc- ture of IP Networks. and addresses some of the limitations of basic shortest path schemes mentioned earlier. This new architecture does nnt need the traditional IP routers tn be replaced or modified. Rather it requires simple devices (such as PCs or network processors) to be carefully placed inside the intradomain network. creating overlay paths between source- destination (SD) pairs. Furthermore. the architecture allows grudirul deployment of such devices. resulting in improved network performance with the addition of each new device. This provides ISPs with an alternative solution to achieve desired level of performance at potentially much lower costs. We will give a brief description of this architecture in Section IV.Keywords
This publication has 15 references indexed in Scilit:
- A localized adaptive proportioning approach to QoS routingIEEE Communications Magazine, 2002
- Resilient overlay networksPublished by Association for Computing Machinery (ACM) ,2001
- Constrained optimization via stochastic approximation with a simultaneous perturbation gradient approximationAutomatica, 1997
- Optimization of discrete event systems via simultaneous perturbation stochastic approximationIIE Transactions, 1997
- Stochastic Optimization by Simulation: Convergence Proofs for the GI/G/1 Queue in Steady-StateManagement Science, 1994
- Multivariate stochastic approximation using a simultaneous perturbation gradient approximationIEEE Transactions on Automatic Control, 1992
- Distributed asynchronous optimal routing in data networksIEEE Transactions on Automatic Control, 1986
- On Approximate Computer System ModelsJournal of the ACM, 1975
- Multidimensional Stochastic Approximation MethodsThe Annals of Mathematical Statistics, 1954
- Stochastic Estimation of the Maximum of a Regression FunctionThe Annals of Mathematical Statistics, 1952