DocumentCode :
2665079
Title :
Hybrid Maximal Matching for Input Buffered Crossbar Switches
Author :
Al Sayeed, Choudhury A. ; Ahmed, Dewan Tanvir ; Rahbar, Akbar Ghaffar Pour
Author_Institution :
Sch. of Inf. Technol. & Eng., Univ. of Ottawa, Ottawa, ON
fYear :
2007
fDate :
14-17 May 2007
Firstpage :
397
Lastpage :
403
Abstract :
To alleviate the problems of network bottleneck in speed networks, it is important to have faster switching and certainly efficient scheduling algorithms. Optical switches are becoming more popular because of their faster switching speeds and higher reliability as compared to their electronic counterpart. Suitable scheduling algorithms are needed to be developed that will improve the utilization of the switch bandwidth, as well as reduce the scheduling overhead. However, most of the present scheduling algorithms try to optimize either the switch bandwidth or the overhead. In this paper, we propose one new scheduling algorithm named as hybrid maximal matching (HMM) that provides a perfect balance between these two performance criteria and proves itself suitable for large size optical switches. The proposed HMM algorithm is simple to implement and arrives with a complexity of at most 0(N) for an NxN input queued crossbar switch.
Keywords :
bandwidth allocation; computational complexity; optical communication equipment; optical fibre networks; optical switches; scheduling; computational complexity; high-speed network; hybrid maximal matching; input buffered crossbar switches; large-size optical switches; scheduling algorithm; switch bandwidth utilization; Bandwidth; Fabrics; Hidden Markov models; Impedance matching; Micromechanical devices; Optical buffering; Optical packet switching; Optical switches; Scheduling algorithm; Throughput;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Communication Networks and Services Research, 2007. CNSR '07. Fifth Annual Conference on
Conference_Location :
Frederlcton, NB
Print_ISBN :
0-7695-2835-X
Type :
conf
DOI :
10.1109/CNSR.2007.36
Filename :
4215541
Link To Document :
بازگشت