On deadlock frequency during dynamic reconfiguration in NOWs

3Citations
Citations of this article
9Readers
Mendeley users who have this article in their library.

This article is free to access.

Abstract

NOWs executing multimedia and real-time applications need to handle dynamic changes in their irregular topologies. This may be carried out in two ways: statically and dynamically. In the former, user traffic is stopped, causing latencies to increase dramatically. In the latter, user traffic is not stopped but deadlocks may appear in the transition phase between the old and the new routing function. To solve this problem, dynamic reconfiguration methods based on deadlock avoidance have been proposed. However, another possibility not studied yet, is to use dynamic reconfiguration allowing deadlock formation with an efficient detection/recovery mechanism. It is necessary to know deadlock frequency during the reconfiguration process in order to assess the viability of this alternative proposal. In this paper, we show that deadlocks may become very infrequent with few virtual channels and the lost message problem can be reduced by using a simple misrouting technique.

Cite

CITATION STYLE

APA

Fernández, L., García, J. M., & Casado, R. (2001). On deadlock frequency during dynamic reconfiguration in NOWs. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 2150, pp. 630–638). Springer Verlag. https://doi.org/10.1007/3-540-44681-8_90

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free