Adaptive timeout discovery using the Network Weather Service

Proceedings 11th IEEE International Symposium on High Performance Distributed Computing - Trang 35-41
M.S. Allen1, R. Wolski1, J.S. Plank2
1Computer Science Department, University of California, Santa Barbara, USA
2Computer Science Department, University of Tennessee, Knoxville, USA

Tóm tắt

In this paper we present a novel methodology for improving the performance and dependability of application-level messaging in Grid systems. Based on the Network Weather Service, our system uses nonparametric statistical forecasts of request-response times to automatically determine message timeouts. By choosing a timeout based on predicted network performance, the methodology improves application and Grid service performance as extraneous and overly-long timeouts are avoided. We describe the technique, the additional execution and programming overhead it introduces, and demonstrate the effectiveness using a wide-area test application.

Từ khóa

#Application software #Weather forecasting #Computer science #Testing #Dynamic scheduling #Degradation #Programming profession #Grid computing #Processor scheduling #Mesh generation

Tài liệu tham khảo

paxon, 1997, Why We Don't Know How to Simulate the Internet

10.1109/90.649563

petitet, 2001, Numerical libraries and the grid, (to appear) Proc of SC01

plank, 2002, The effect of timeout prediction and selection on wide area collective operations, IEEE Int l Symp Network Computing Applications

spring, 1998, Application level scheduling: Gene sequence library comparison, Proceedings of the 1998 ACM International Conference on Supercomputing

10.1109/INFCOM.2000.832571

10.1023/A:1019025230054

10.1016/S0167-739X(99)00025-4

forum, 1994, Mpi: A message-passing interface standard, Technical Report CS-94–230

10.1109/90.793002

10.1155/2000/319291

10.1109/HCW.2000.843757

kielmann, 1999, Magpie: Mpi's collective communication operations for clustered wide area systems, Proc ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming (PPoPP'99), 131

10.1016/S0167-8191(01)00098-9

berman, 1996, Application level scheduling on distributed heterogeneous networks, Proceedings of Supercomputing 1996, 10.1145/369028.369109

berman, 2000, The GrADS project: Software support for high-level grid application development, Technical Report Rice COMPTR00-355

10.1109/IPPS.1996.508087

Thông tin
Thông tin xuất bản

Proceedings 11th IEEE International Symposium on High Performance Distributed Computing

35-41

Thông tin tác giả