A unified framework for cluster manager election and clustering mechanism in mobile ad hoc networks
Introduction
With the explosion of wireless network, the activity of mankind has been impacted, in the infrastructure network structure; the medium for fixing center will be regarded as the relay station to bringing about the connection of two mobile equipments, so that the data is transmitted. The relay station, such as an access point, was given an excellent feature to control the position of mobile equipments, which is just the same as the router. But these equipments are usually damaged by the extrinsic factors such as war, natural disasters, etc., which will cause the failure connection between the Mobile Nodes (MNs), therefore, the traditional wireless network has failed to meet the needs of mankind. In recent years, many experts and scholars have started paying attention to the non-infrastructure network structure, in which, the Mobile Ad-hoc NETwork's (MANET) purpose is to have the freedom to gain information in anytime and anyplace under a non-infrastructure network structure for meeting the needs of mankind. Therefore, MANET should pay more attention to all fields of application. In MANET, there are typically four kinds of applications such as the private local area network, military environment, civil and emergency, etc. [7].
The strong points of MANET are that it is easy and fast to deploy, which does not need the reply from the fixed relay station to quickly log on in anytime and anyplace. Therefore, there is a research of topic under MANET in this paper. The researched MANET consists of many mobile nodes such as the mobile phone, portable computer, PDA, etc.
At present, almost wireless network structure must depend on the base station of connecting with the bone via the wired system, which means this network structure is still fixed and limed to the signal of base station, while it must cost a lot of money to build this wireless network environment, with the difficulty to finish the building at once when the emergency happens. Therefore, MANET is the peer-to-peer network, under the situation of failing to connect with the bone network in the battlefield and natural disaster; MANET will be playing an important role. This network structure can make every mobile node move; the signal will be transmitted via the delivery between the mobile nodes.
MANET consists of many Cluster Managers (CMs) and mobile nodes. The given node of cluster manager centers on forming the clustering and keeping the network topology. CM will make the resource allocation and control supervision for all nodes in the relevant clustering. The other nodes except for the manager are all named as the mobile nodes. Under this structure, the nodes need not the base station to dynamically form the network with each other. MN can be free to move and connect with each other via multi-hop wireless links; every MN can play the role of router to help the other nodes of network transmitting the data package. Due to the features of MANET, its structure is very usable in the specific environments such as the rescue of the battlefield and rescue of disaster, the rebuilding after the earthquake, the discovery to some unknown area, etc. However, just because MN can be free to move everywhere, the network topology appears the dynamic change.
With the instant development of network, the connecting type of network model is various, the research on the clustering and election of manager in [2], [3], [6], [10], [15], [16] did not consider the operating mode and building of backup manager under the initial stage of network and finished group. Every mobile node has the characteristic of dynamic action, therefore, the nodes will continue to be clustered and decomposed, which will affect the stability of network topology, and cause constantly the reconfiguration of the network groups, all that will influence the overall function of network. Therefore, it is an important issue on how to choose a CM to keep the stability of network topology. The election of suitable CM is NP hard [3], the existing method for solving this kind of problems almost is based on serving as heuristic, without the relevant research trying to keep the stability of network topology [3], [6]. However, a good clustering network structure should keep the stability of structure as possible when the node is moving or the network topology is changed, otherwise, the redistribution of groups and the transfer of the numerous information will cause the network burden and the cost raising. Because all things are particular about the reliability and usability today, the building of a mechanism should consider the various situations. The passing research failed to consider the initial stage of network and finished clustering, so when the manager fails, the needed redistribution of groups would influence the efficiency and stability. Therefore, this technology will research the method of clustering and election of managers according to the different network nodes in MANET, and add the mechanism of backup manager to reduce the disadvantages and improve the efficiency when the original manager does not work.
This research serves as MANET to design Unified Framework Clustering Mechanism (UFCM for short) for the backup manager, which centers on how does MANET bring the clustering and election of manager under the limited conditions (transmission speed and the quantity of mobile nodes), with the least times of data exchange and network transmission.
The second part of this paper will treat the issue concerned; the third part introduced the method and frame of this research; the fourth part is the test result; the fifth part is the conclusion and the future works of research.
Section snippets
Literature review
As MANET need not any infrastructure to provide the multi-hop wireless links for the mobile user, the network will offer the mechanism for the simultaneous uses of many users in order to widely apply for the field of actual practice. However, the method of search-address and ringing is more difficult than the common network. For this reason, the hierarchical search-address of MANET is able to efficiently solve the problems of complex search-address, while clustering is used for setting up and
The proposed method
Generally, under the real environment, all mobile nodes are possible to be distributed in the different place, so they need to communicate and coordinate with each other; such environment is named as the environment of MANET. Under such circumstance, it is very important to know how to efficiently cluster and select manager.
There are many small networks in MANET, and every small network is given a lot of mobile nodes, this research is purposed to make every mobile node finish the election of
Simulation experiment
There is a compilation on the actual procedures as the method of clustering of this research according to the method in the third part. In the experiment, the environment of topology is designated as an area of 1000 ⁎ 1000, the nature of every node consists of Node_ID, Pos_X, Pos_Y, Direction, Speed, Range, Energy, Busy, and Bandwidth, of which the parameters will produce 1000 nodes and the relevant nature value under NS-2 [21] through the random value producer except the Node_ID, while it is
Conclusion and future works
There is an applicable treatment of relevant issue on MANET in this research. According to the researches in this field from the past scholars, their many methods on the election of manager and clustering in MANET did not consider the different stage of network, not develop the suitable method on the election of manager and clustering either, even failed to consider the reason why the manager fails.
For these reasons, this research brings a treatment of the issue on the mechanism of election of
Shu-Ching Wang received her B.S. degree in Computer Science from Feng-Chia University, M.S. degree in Electrical Engineering from National Chen-Kung University, and Ph.D. degree in Information Engineering from National Chiao-Tung University, Taiwan.
Currently, she is a Professor with the Graduate Institute of Informatics, Chaoyang University of Technology, Taiwan. Her current research interests include distributed data processing, parallel processing, algorithm analysis and design, mobile
References (21)
- et al.
“The architectural organization of a mobile radio network via a distributed algorithm,”
IEEE Transactions on Communications
(1981) “Distributed clustering for ad hoc networks,”
- et al.
“A generalized clustering algorithm for peer-to-peer networks,”
“Accomplishments of the DARPA survivable adaptive networks SURAN program,”
- et al.
“Connectivity-based k-hop clustering in wireless networks,”
- et al.
“A new approach to the design and analysis of peer-to-peer mobile networks,”
Wireless Networks
(1999) - Y.Y. Chou, “A Study of the Hierarchical Cellular-Based Management Worked by Power-Aware Technology”, Master's thesis,...
- et al.
“Routing in ad-hoc networks using minimum connected dominating sets,”
- et al.
“Fast distributed algorithms for (weakly) connected dominating sets and linear-size skeletons,”
- et al.
“Multicluster, mobile, multimedia radio network,”
Wireless Networks
(1995)
Cited by (0)
Shu-Ching Wang received her B.S. degree in Computer Science from Feng-Chia University, M.S. degree in Electrical Engineering from National Chen-Kung University, and Ph.D. degree in Information Engineering from National Chiao-Tung University, Taiwan.
Currently, she is a Professor with the Graduate Institute of Informatics, Chaoyang University of Technology, Taiwan. Her current research interests include distributed data processing, parallel processing, algorithm analysis and design, mobile computing, and grid computing.
Hsin-Hung Pan received his B.S. degree in Industrial Engineering and Management and M.S. degree in Information Management from Chaoyang University of Technology, Taiwan.
Currently, he is pursuing his Ph.D. in Graduate Institute of Informatics, Chaoyang University of Technology, Taiwan. His current research interests include distributed data processing, fault tolerant computing, and mobile computing.
Kuo-Qin Yan received his B.S. and M.S. degrees in Electrical Engineering from Chung Cheng Institute of Technology and his Ph.D. degree in Computer Sciences from National Tsing-Hua University, Taiwan.
Currently, he is a Professor with the Department of Business Administration, Chaoyang University of Technology, Taiwan. His current research interests include distributed data processing, parallel processing, fault tolerant computing, mobile computing, and grid computing.
Yu-Lung Lo received his M.S. and Ph.D. degrees in Computer Sciences from the University of Central Florida, USA.
Currently, he is an Associate Professor with the Department of Information Management, Chaoyang University of Technology, Taiwan. His current research interests include distributed data processing, parallel processing, mobile computing, and multimedia database system.