TY - JOUR
T1 - Multicast routing with multiple QoS constraints in ATM networks
AU - Wu, Jang Jiin
AU - Hwang, Ren Hung
AU - Lu, Hsueh I.
PY - 2000/5
Y1 - 2000/5
N2 - Recently, more and more applications provide multiparty communication services, e.g., video conferencing, distance learning, etc. Therefore, the routing problem of multicast with multiple quality of service (QoS) constraints becomes more important. For example, in order to ensure smooth play back of audio and video data, a video conference requires guarantee on both end-to-end delay and loss probability. In this paper, we solve the multiple-constraint multicast problem by extending three single-constraint Steiner tree algorithms: (1) the constrained shortest path tree (CSPT), (2) the algorithm proposed by Kompella, Pasquale, and Polyzos (KPP), and (3) the bounded shortest multicast algorithm (BSMA). We also propose a novel multiple-constraint multicast routing algorithm based on genetic algorithms, called MCMGA. These four algorithms are evaluated via simulations on different sizes of random graphs. Our numerical results show that, in most cases, MCMGA yields solutions of the least cost. The costs of the solutions obtained by our extended BSMA are also very competitive.
AB - Recently, more and more applications provide multiparty communication services, e.g., video conferencing, distance learning, etc. Therefore, the routing problem of multicast with multiple quality of service (QoS) constraints becomes more important. For example, in order to ensure smooth play back of audio and video data, a video conference requires guarantee on both end-to-end delay and loss probability. In this paper, we solve the multiple-constraint multicast problem by extending three single-constraint Steiner tree algorithms: (1) the constrained shortest path tree (CSPT), (2) the algorithm proposed by Kompella, Pasquale, and Polyzos (KPP), and (3) the bounded shortest multicast algorithm (BSMA). We also propose a novel multiple-constraint multicast routing algorithm based on genetic algorithms, called MCMGA. These four algorithms are evaluated via simulations on different sizes of random graphs. Our numerical results show that, in most cases, MCMGA yields solutions of the least cost. The costs of the solutions obtained by our extended BSMA are also very competitive.
UR - http://www.scopus.com/inward/record.url?scp=0033900102&partnerID=8YFLogxK
U2 - 10.1016/S0020-0255(99)00102-4
DO - 10.1016/S0020-0255(99)00102-4
M3 - Article
AN - SCOPUS:0033900102
SN - 0020-0255
VL - 124
SP - 29
EP - 57
JO - Information sciences
JF - Information sciences
IS - 1
ER -