| F. Petrini and W. chun Feng. Scheduling with global information in distributed systems. In 20th International Conference on Distributed Computing Systems (ICDCS 2000. |
....on three small examples. Section 4 presents some of the ideas generated from the backgroud study and from the case study experiments. Finally in Section 5 a general description of the program implementing the proposed algorithms is given. 2 Background At the beginning of the stage several papers [9, 18, 17, 25, 12, 4, 5, 6, 20, 23, 22, 21, 2, 1, 14, 7, 15, 24, 19] have been studied to obtain the necessary background. In particular reviewing Dima s paper [8] proved essential for the remainder of the work and also yielded some suggestions for improvement to the author. pinello eecs.berkeley.edu 1 http: www rocq.inria.fr syndex 2 ....
F. Petrini and W. chun Feng, Scheduling with global information in distributed systems, in Proc. 20th IEEE International Conference on Distributed Computing Systems, April 2000.
No context found.
Fabrizio Petrini and Wu-chun Feng. Scheduling with Global Information in Distributed Systems. In Proceedings of the The 20th International Conference on Distributed Computing Systems, Taipei, Taiwan, Republic of China, April 2000.
....do not intend at this stage to modify the OS kernel, so the comparison between di erent local scheduling policies will be con ned to the operating systems available for our testing. 2.3 Bu ered Coscheduling 2.3. 1 Overview Bu ered coscheduling (BCS) represents a di erent approach to coscheduling [10, 11, 12, 13, 21]. In BCS, local scheduling decisions are based on global information of the sys8 tem s status, essentially converting an on line problem of coordinating jobs to an o ine problem. Global knowledge of the system s status can potentially be costly and unscalable, so BCS o ers a technique for seamless ....
Fabrizio Petrini and Wu-chun Feng. Scheduling with Global Information in Distributed Systems. In Proceedings of the The 20th International Conference on Distributed Computing Systems, Taipei, Taiwan, Republic of China, April 2000.
....the low level features of the interconnection network. For example, if control packets are given higher priority than background traffic at the sending and receiving endpoints, they can be delivered with predictable network latency 2 during the execution of a direct total exchange algorithm 3 [14]. The global knowledge of the communication pattern provided by the total exchange allows for the implementation of efficient flow control strategies. For example, it is possible to avoid congestion inside the network by carefully scheduling the communication pattern and limiting the negative ....
F. Petrini and W. Feng. Scheduling with Global Information in Distributed Systems. In Proceedings of the International Conference on Distributed Computing Systems (ICDCS'00), April 2000.
No context found.
F. Petrini and W. chun Feng. Scheduling with global information in distributed systems. In 20th International Conference on Distributed Computing Systems (ICDCS 2000.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC