A Distributed Service Management Infrastructure for Enterprise Data Centers Based on Peer-to-Peer Technology

Authors:
Chunqiang Tang;Rong N. Chang;Edward So
Affiliations:
IBM Thomas J. Watson Research Center;IBM Thomas J. Watson Research Center;IBM Thomas J. Watson Research Center
Venue:
SCC '06 Proceedings of the IEEE International Conference on Services Computing
Year:
2006

Citing 0
Cited 6

A scalable application placement controller for enterprise data centers

Proceedings of the 16th international conference on World Wide Web
Dynamic quorums for DHT-based enterprise infrastructures

Journal of Parallel and Distributed Computing
Harnessing the power of DHTs to build dynamic quorums in large-scale enterprise infrastructures

LADIS '08 Proceedings of the 2nd Workshop on Large-Scale Distributed Systems and Middleware
A latency-aware algorithm for dynamic service placement in large-scale overlays

IM'09 Proceedings of the 11th IFIP/IEEE international conference on Symposium on Integrated Network Management
A2A: An Architecture for Autonomic Management Coordination

DSOM '09 Proceedings of the 20th IFIP/IEEE International Workshop on Distributed Systems: Operations and Management: Integrated Management of Systems, Services, Processes and People in IT
Bulletin board: a scalable and robust eventually consistent shared memory over a peer-to-peer overlay

ACM SIGOPS Operating Systems Review

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper presents a distributed service management infrastructure called BISE. One distinguishing feature of BISE is its adoption of the Peer-to-Peer (P2P) model in support of realtime service managements. BISE offers significant advantages over existing systems in scalability, resilience, and manageability. Current P2P algorithms are mainly developed for the file-sharing applications running on desktops, which have characteristics dramatically different from enterprise data centers. This difference led us to design our own P2P algorithms specifically optimized for enterprise environments. Based on these algorithms, we implemented a P2P substrate called BiseWeaver (25,000 lines of Java code) as the core of BISE. Our evaluation on a set of distributed machines shows that BiseWeaver is efficient and robust, and provides timely monitoring data in support of proactive SLA management.