Traffic engineering for provisioning restorable hose-model VPNs
Resource
Next Generation Internet Design and Engineering, 2006. NGI '06. 2006 2nd Conference on
Journal
Next Generation Internet Design and Engineering, 2006. NGI '06. 2006 2nd Conference on
Pages
-
Date Issued
2006-04
Date
2006-04
Author(s)
DOI
N/A
Abstract
Virtual private networks (VPNs) are overlay networks established on top of a public network backbone with the goal of providing a service comparable to private networks (PNs). The recently proposed VPN hose-model provides customers with flexible and convenient ways to specify their bandwidth requirements. In order to meet specified bandwidth requirements, the network service provider (NSP) must reserve a sufficient amount of bandwidth on the data transmission paths between each endpoint pair of a VPN. In addition, reliability of a VPN depends on the reliability of data transmission paths. Italiano et al. proposed an algorithm that finds a set of backup paths for a given VPN (VPN tree) under the single-link failure model. When any link failure on the VPN tree is detected, a backup path corresponding to the failed link can be activated to restore the disconnected VPN tree into a new one, and hence can ensure reliability of the given VPN. However, Italiano's algorithm cannot guarantee the specified bandwidth requirement of the given VPN under the single-link failure model will be met. To address this issue a new backup path set selection algorithm called BANGUAD is proposed in this work. In addition, issues about establishing multiple bandwidth-guaranteed hose-model VPNs under the single-link failure model have not been investigated. In this paper, we propose a bandwidth sharing algorithm as well as three provisioning algorithms for establishing multiple bandwidth-guaranteed hose-model VPNs under the single-link failure model. Experimental simulations that compare the performance of the proposed algorithms are reported.
SDGs
Type
other
File(s)![Thumbnail Image]()
Loading...
Name
01678234.pdf
Size
312.75 KB
Format
Adobe PDF
Checksum
(MD5):91058b483dbe09df9e4cafcc3389bbaf