DocumentCode
2418865
Title
Efficient multicast in heterogeneous networks of workstations
Author
Libeskind-Hadas, Ran ; Line, Jeffrey Hart
Author_Institution
Dept. of Comput. Sci., Harvey Mudd Coll., Claremont, CA, USA
fYear
2000
fDate
2000
Firstpage
403
Lastpage
410
Abstract
This paper studies the problem of efficient multicast in heterogeneous networks of workstations (HNOWs) using a parameterized communication model. This model associates a sending overhead and a receiving overhead with each node as well as a network latency parameter. The problem of finding optimal multicasts in this model is known to be NP-complete in the strong sense. Nevertheless, we show that for two different properties that arise in typical HNOWs, provably near-optimal and optimal solutions, respectively, can be found in polynomial time. Specifically, we show the following two results: when the ratios of receiving overhead to sending overhead among the nodes is bounded by constants, solutions within a bounded ratio of optimal can be found in time O(nlogn). Secondly, if the number of distinct types of workstations is fixed then optimal solutions can be found in polynomial time. These results provide a practical means of finding optimal and provably near-optimal multicast schedules in a large class of frequently occurring heterogeneous networks of workstations
Keywords
communication complexity; multicast communication; workstation clusters; efficient multicast; heterogeneous workstation networks; network latency parameter; optimal multicasts; parameterized communication model; receiving overhead; sending overhead; Computer science; Costs; Delay; Educational institutions; Intelligent networks; Multicast communication; Polynomials; Processor scheduling; Radio access networks; Workstations;
fLanguage
English
Publisher
ieee
Conference_Titel
Parallel Processing, 2000. Proceedings. 2000 International Workshops on
Conference_Location
Toronto, Ont.
ISSN
1530-2016
Print_ISBN
0-7695-0771-9
Type
conf
DOI
10.1109/ICPPW.2000.869145
Filename
869145
Link To Document