A fast parallel algorithm for the maximal independent set problem
Journal of the ACM (JACM)
A simple parallel algorithm for the maximal independent set problem
SIAM Journal on Computing
Constructing a maximal independent set in parallel
SIAM Journal on Discrete Mathematics
Strictly nonblocking f-cast photonic networks
IEEE/ACM Transactions on Networking (TON)
Light trees: optical multicasting for improved performance in wavelength routed networks
IEEE Communications Magazine
A large scalable ATM multicast switch
IEEE Journal on Selected Areas in Communications
Power-efficient design of multicast wavelength-routed networks
IEEE Journal on Selected Areas in Communications
Hi-index | 0.00 |
This paper investigates the throughput performance of the d-tree multicast capable optical cross-connect (MC-OXC) by using the separated unicast/multicast splitter-and-delivery (SUM-SaD) switches under synchronous traffic. A d-tree property is introduced to constrain the ability that the MC-OXC and SUM-SaD can accommodate the numbers of light trees. We propose the maximal independent set (MIS) model of the conflict graph to establish as most sessions as possible. It needs to find the d- constrained MIS (d-MIS) of multicast conflict graph and the MIS of the remained unicast conflict graph to maximize the throughput under the assumption that the multicast sessions have higher priority than the unicast ones. The MIS problem of the conflict graph is NP-complete, and two heuristic conflict-based algorithms: Minimal Conflict First (MCF) and First See First Service (FSFS) are proposed to find the MIS of the conflict graph. The Minimal Member First (MMF) and Maximal Member First (MAMF) algorithms are applied to determine the d-MIS of multicast conflict graph uniquely. Simulation results show that the d-tree MC-OXC at d ≥ 2 achieves favorable throughput performance. When the wavelength resource on each fiber link is scarce, the performances of the four conflict-based MIS algorithms are different and behave in the order of MCF, MMF, FSFS and MAMF from best to worst.