loading...
 This Article 
   
 Share 
   
 Bibliographic References 
   
 Add to: 
 
Digg
Furl
Spurl
Blink
Simpy
Google
Del.icio.us
Y!MyWeb
 
 Search 
   
Fourth IEEE International Symposium on Cluster Computing and the Grid (CCGrid'04)
GTP: group transport protocol for lambda-Grids
Chicago, IL, USA
April 19-April 22
ISBN: 0-7803-8430-X
R.X. Wu, Dept. of Comput. Sci. & Eng., Univ. of California, San Diego, CA, USA
A.A. Chien, Dept. of Comput. Sci. & Eng., Univ. of California, San Diego, CA, USA
The notion of lambda-Grids posits plentiful collections of computing and storage resources richly interconnected by dedicated dense wavelength division multiplexing (DWDM) optical paths. In lambda-Grids, the DWDM links form a network with plentiful bandwidth, pushing contention and sharing bottlenecks to the end systems (or their network links) and motivating the group transport protocol (GTP). GTP features a request-response data transfer model, rate-based explicit flow control, and more importantly, receiver-centric max-min fair rate allocation across multiple flows to support multipoint-to-point data movement. Our studies show that GTP performs as well as other UDP based aggressive transport protocols (e.g. RBUDP, SABUL) for single flows, and when converging flows (from multiple senders to one receiver) are introduced, GTP achieves both high throughput and much lower loss rates than others. This superior performance is due to new techniques in GTP for managing end system contention.
Citation:
R.X. Wu, A.A. Chien, "GTP: group transport protocol for lambda-Grids," ccgrid, pp.228-238, Fourth IEEE International Symposium on Cluster Computing and the Grid (CCGrid'04), 2004
Usage of this product signifies your acceptance of the Terms of Use.