Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
A fault-tolerant group communication protocol in large scale and highly dynamic mobile next-generation networks
Cao J., Wang G., Chan K. IEEE Transactions on Computers56 (1):80-94,2007.Type:Article
Date Reviewed: Jul 30 2007

Mobile next-generation networks (NGNs) that integrate mobile and wireless networks with the wired Internet are being widely deployed, and services that enable users to do group-oriented mobile collaborations, where messages from one or multiple sources are delivered to a large number of recipients, are gaining in popularity. The RingNet architecture and fault-tolerant group communication protocol proposed in this paper is novel in that it addresses the reliability issues in mobile NGNs.

The paper presents an overview of current research protocols to support group communication in the wired Internet and mobile networks. These protocols often use some form of tree-based hierarchy to address the scalability requirement. However, such networks have a single point of failure: when a node in the hierarchy becomes faulty, the tree is broken into several subtrees that need to reconfigure into a tree. A common shortcoming of existing approaches is that, while they are designed to work with message loss due to unreliable communication links, they do not consider node failures.

The key contribution of this paper is a network architecture that handles node failures. The proposed RingNet architecture is a hierarchy that is a combination of logical rings and logical trees. The nodes in the network are organized as logical trees, providing the scalability needed for a group communication protocol. Each node also belongs to a logical ring, so that when a node fails, the resulting subtrees can heal quickly, namely recreate the tree hierarchy via sibling nodes in the logical ring.

The main tasks that must be supported by a group communication protocol are, first, group membership management to maintain information regarding member nodes that may voluntarily join or leave a group, or cease to be a member due to failure, and, second, group multicast dissemination to efficiently deliver information from one or more sources to all the member nodes in the group. The paper presents algorithms to support the group membership maintenance task in the RingNet architecture. These consist of a membership-propagation algorithm that can detect several types of failures, and a topology-maintenance algorithm with ring-repair and hierarchy-repair procedures.

A theoretical analysis is used to prove that the RingNet hierarchy has the same scalability, but is more reliable than a tree-based hierarchy. Two important properties of the proposed fault-tolerant topology-maintenance algorithm, the timeliness property, which quantifies how fast the protocol responds to failure events, and the accuracy property, which quantifies how well the protocol maintains the hierarchy in the presence of node failures, are validated in average cases via simulations using the ns-2 simulation tool.

As group-oriented applications such as news feeds and streaming live audio and video gain in popularity, group communication protocols that can handle the issues of mobile NGNs become increasingly important. This paper presents a validated solution that addresses a key requirement, reliability in the presence of node failures, and is relevant for researchers investigating group communication protocols.

Reviewer:  Suma Adabala Review #: CR134581 (0807-0677)
Bookmark and Share
  Editor Recommended
 
 
Network Protocols (C.2.2 )
 
 
Fault Tolerance (C.4 ... )
 
 
Network Management (C.2.3 ... )
 
 
Reliability, Availability, And Serviceability (C.4 ... )
 
 
Wireless Communication (C.2.1 ... )
 
 
Network Operations (C.2.3 )
 
Would you recommend this review?
yes
no
Other reviews under "Network Protocols": Date
An implementation of an automated protocol synthesizer (APS) and its application to the X.21 protocol
Ramamoorthy C. (ed), Dong S., Usuda Y. IEEE Transactions on Software Engineering SE-11(9): 886-908, 1985. Type: Article
Apr 1 1986
From state machines to temporal logic: specification methods for protocol standards
Schwartz R., Melliar-Smith P.  The analysis of concurrent systems (, Cambridge,651985. Type: Proceedings
Aug 1 1986
Reaching approximate agreement in the presence of faults
Dolev D., Lynch N., Pinter S., Stark E., Weihl W. Journal of the ACM 33(3): 499-516, 1986. Type: Article
Aug 1 1988
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy