MPLS Label Stacking on the Line Network

Abstract

Abstract. All-Optical Label Switching (AOLS) is a new technology that performs forwarding without any Optical-Electrical-Optical conversions. The most promising scheme to manage the control plane of these optical networks is Generic MultiProtocol Label Switching (GMPLS). In this paper, we study the problem of routing a set of requests in GMPLS networks with the aim of minimizing the number of labels required to ensure the forwarding. In order to spare the label space, we consider label stacking, allowing the configuration of GMPLS tunnels. We study particularly this network design problem when the network is a line. We provide an exact algorithm for the case in which all the requests have a common source and present some approximation algorithms and heuristics when an arbitrary number of sources are distributed over the line. We analyze by simulations the performance of our proposed algorithms and compare them with previous ones.

Designing Hypergraph Layouts to GMPLS Routing Strategies

“…For the sake of presentation, we describe our algorithms when the network is a path, a tree, and a general graph in Sections 4.1, 4.2, and 4.3, respectively. The approximation algorithm for the directed path network appeared also in [2], we include it here for the sake of completeness.…”

“…We summarize the algorithm that appeared first in [2] (Gerstel et al used a similar approach in ??). In the case of a single source, it is not difficult to see that the tunnel structure is non crossing, i.e.…”

“…In the case of a single source, it is not difficult to see that the tunnel structure is non crossing, i.e. two tunnels can only intersect in an optimal solution if one is strictly inside the other [2]. Since the path is directed, we assume w.l.o.g.…”

“…For instance, in [16] the authors propose heuristics for routing a set of demands in AOLS networks when routers have limited number of available optical correlators. Very recently [2], the problem has been studied for the directed path from a more theoretical point of view. Namely, in [2] the authors present a polynomial-time optimal algorithm for the case when all traffic is issued from a single source and an O(log n)-approximation algorithm with arbitrary number of sources, where n is the number of nodes of the network.…”

“…The approximation algorithms strongly rely on the already known algorithms for VPL problems. Finally, we focus on the path topology, extending the dynamic programming approach presented in [2] to any bounded number of sources. If there are k sources, the main result is an optimal algorithm with running time n O(k) .…”