Randomized gossip algorithms for maintaining a distributed bulletin board with guaranteed age properties
Abstract
Scalable computer systems, including clusters and multi‐cluster grids, require routine exchange of information about the state of system‐wide resources among their nodes. Gossip‐based algorithms are popular for providing such information services due to their simplicity, fault tolerance and low communication overhead. This paper presents a randomized gossip algorithm for maintaining a distributed bulletin board among the nodes of a scalable computer system. In this algorithm each node routinely disseminates its most recently acquired information while maintaining a snapshot of the other nodes' states. The paper provides analytical approximations for the expected average age, the age distribution and the expected maximal age for the acquired information at each node. We confirm our results by measurements of the performance of the algorithm on a multi‐cluster campus grid with 256 nodes and by simulations of configurations with up to 2048 nodes. The paper then presents practical enhancements of the algorithm, which makes it more suitable for a real system. Such enhancements include using fixed‐size messages, reducing the number of messages sent to inactive nodes and supporting urgent information. The enhanced algorithm guarantees the age properties of the information at each node in the configurations with an arbitrary number of inactive nodes. It is being used in our campus grid for resource discovery, for dynamic assignment of processes to the best available nodes, for load‐balancing and for on‐line monitoring. Copyright © 2009 John Wiley & Sons, Ltd.
Number of times cited: 7
-
Vishakha Gupta-Cledat,
Donald E. Porter and Vivek
Sarkar
the12th ACM SIGPLAN/SIGOPS International Conference
VEE '16
Atlanta, Georgia, USA
Proceedings of the12th ACM SIGPLAN/SIGOPS International Conference on Virtual Execution
Environments - VEE '16
Virtual Execution Environments
ACM Press
New York, New York, USA
, (2016).
9781450339476
, 10.1145/2892242
20160328091845
http://dl.acm.org/citation.cfm?doid=2892242
Muli Ben-Yehuda, Orna Agmon Ben-Yehuda and Dan
Tsafrir
The nom Profit-Maximizing Operating System
, (2016).
145
160
, 10.1145/2892242.2892250
20160328091845
http://dl.acm.org/citation.cfm?doid=2892242.2892250
- Philip Healy, Theo Lynn, Enda Barrett and John P. Morrison, Single system image: A survey, Journal of Parallel and Distributed Computing, 10.1016/j.jpdc.2016.01.004, 90-91, (35-51), (2016).
- Muli Ben-Yehuda, Orna Agmon Ben-Yehuda and Dan Tsafrir, The nom Profit-Maximizing Operating System, ACM SIGPLAN Notices, 10.1145/3007611.2892250, 51, 7, (145-160), (2016).
- Amnon Barak, Zvi Drezner, Ely Levy, Matthias Lieber and Amnon Shiloh, Resilient gossip algorithms for collecting online management information in exascale clusters, Concurrency and Computation: Practice and Experience, 27, 17, (4797-4818), (2015).
- Pablo Chacin, Leando Navarro and Pedro Garcia Lopez, Utility Driven Service Routing over Large Scale Infrastructures, Towards a Service-Based Internet, 10.1007/978-3-642-17694-4_8, (88-99), (2010). Seventh IEEE International Symposium on Cluster Computing and the Grid Rio de Janeiro, Brazil Seventh IEEE International Symposium on Cluster Computing and the Grid (CCGrid '07) IEEE , (2007). 0-7695-2833-3 Lior Amar, Amnon Barak, Ely Levy and Michael Okun An On-line Algorithm for Fair-Share Node Allocations in a Cluster , (2007). 83 91 4215369 , 10.1109/CCGRID.2007.22 http://ieeexplore.ieee.org/document/4215369/
- Başak Altan and Okan Örsan Özener, Cost allocation mechanisms in a peer‐to‐peer network, Networks, , (2018).
