Optimal bandwidth reservation in hose-model VPNs with multi-path routing

A virtual private network (VPN) provides private network connections over a publicly accessible shared network. Bandwidth provisioning for VPNs leads to challenging optimization problems. In the hose model proposed by Duffield et al., each VPN endpoint specifies bounds on the total amount of traffic that it will send or receive at any time. The network provider must provision the VPN so that there is sufficient bandwidth for any traffic matrix that is consistent with these bounds. While previous work has considered tree routing and single-path routing between the VPN endpoints, we demonstrate that the use of multipath routing offers significant advantages. On the one band, we present an optimal polynomial-time algorithm that computes a bandwidth reservation of minimum cost using multi-path routing. This is in contrast to tree routing and single-path routing, where the problem is computationally hard. On the other hand, we present experimental results showing that the reservation cost using multi-path routing can indeed be significantly smaller than with tree or single-path routing.

[1]  Alexander Schrijver,et al.  Theory of linear and integer programming , 1986, Wiley-Interscience series in discrete mathematics and optimization.

[2]  Kurt Mehlhorn,et al.  A platform for combinatorial and geometric computing , 1995 .

[3]  Amit Kumar,et al.  Algorithms for provisioning virtual private networks in the hose model , 2002, TNET.

[4]  L. Lovász,et al.  Geometric Algorithms and Combinatorial Optimization , 1981 .

[5]  Ravindra K. Ahuja,et al.  Network Flows: Theory, Algorithms, and Applications , 1993 .

[6]  Tim Roughgarden,et al.  Simpler and better approximation algorithms for network design , 2003, STOC '03.

[7]  Albert G. Greenberg,et al.  Resource management with hoses: point-to-cloud services for virtual private networks , 2002, TNET.

[8]  Amit Kumar,et al.  Provisioning a virtual private network: a network design problem for multicommodity flow , 2001, STOC '01.

[9]  István Szabó,et al.  On bandwidth efficiency of the hose resource management model in virtual private networks , 2003, IEEE INFOCOM 2003. Twenty-second Annual Joint Conference of the IEEE Computer and Communications Societies (IEEE Cat. No.03CH37428).

[10]  Prabhakar Raghavan,et al.  Randomized rounding: A technique for provably good algorithms and algorithmic proofs , 1985, Comb..

[11]  Chaitanya Swamy,et al.  Primal–Dual Algorithms for Connected Facility Location Problems , 2004, Algorithmica.

[12]  Rajeev Rastogi,et al.  Restoration algorithms for virtual private networks in the hose model , 2002, Proceedings.Twenty-First Annual Joint Conference of the IEEE Computer and Communications Societies.

[13]  Albert G. Greenberg,et al.  A flexible model for resource management in virtual private networks , 1999, SIGCOMM '99.

[14]  Thomas Erlebach,et al.  Price-based call admission control in a single DiffServ domain , 2002 .

[15]  Daniel O. Awduche,et al.  Requirements for Traffic Engineering Over MPLS , 1999, RFC.