Is Light-Tree Structure Optimal for Multicast Routing in Sparse Light Splitting WDM Networks?

TL;DR

This paper proposes a light-hierarchy structure for multicast routing in sparse light splitting WDM networks, demonstrating improved efficiency and performance over traditional light-trees.

Contribution

It introduces a graph renewal and distance priority algorithm for constructing light-hierarchies, enhancing multicast routing efficiency in WDM networks.

Findings

Light-hierarchies reduce link stress and improve network throughput.

The proposed algorithms outperform existing light-tree methods.

Light-hierarchies achieve lower end-to-end delay and network diameter.

Abstract

To minimize the number of wavelengths required by a multicast session in sparse light splitting wavelength division multiplexing (WDM) networks, a light-hierarchy structure, which occupies the same wavelength on all links, is proposed to span as many destinations as possible. Different from a light-tree, a light-hierarchy accepts cycles, which are used to traverse crosswise a 4-degree (or above) multicast incapable (MI) node twice (or above) and switch two light signals on the same wavelengths to two destinations in the same multicast session. In this paper, firstly, a graph renewal and distance priority light-tree algorithm (GRDP-LT) is introduced to improve the quality of light-trees built for a multicast request. Then, it is extended to compute light-hierarchies. Obtained numerical results demonstrate the GRDP-LT light-trees can achieve a much lower links stress, better wavelength channel cost, and smaller average end-to-end delay as well as diameter than the currently most efficient algorithm. Furthermore, compared to light-trees, the performance in terms of link stress and network throughput is greatly improved again by employing the light-hierarchy, while consuming the same amount of wavelength channel cost.