Efficient self-organized backbone formation in mobile ad hoc networks (MANETs)☆
Graphical abstract
Introduction
Quality of Service in an ad hoc environment has been widely recognized as a challenging problem. Characteristics of ad hoc networks such as mobility, dynamics of the environment and uncertainty of resource availability makes provisioning of QoS guarantees difficult. The performance of MANET categorized by two types based on quality of service: (1) Statefull – each node maintains state information about the traffic flow and connectivity. (2) Stateless – does not maintain the state information, which is scalable in wireless networks. The proposed work self-organized backbone also fall under partially statefull QoS type i.e. only the backbone node maintains state information based on demand. State information’s are maintained by the backbones during the request made by other nodes i.e. a node wish to connect or leave the backbone or require high bandwidth link.
Many routing protocols are developed in the last two decades, all the routing mechanisms are based on the assumption that each node receives the information in one interface and forwards to other interface. Only the intermediate nodes perform these functions, where as the origin or source nodes perform the route finding process in the initial stage. Hence source and destination must rely on intermediate nodes for route finding, data transfer and acknowledgement process. These three functions are involved in the basic packet switching methods like virtual circuit and data gram approaches. From this point of view priority or importance must be given to intermediate nodes, and the routing protocol design should concentrate on these nodes. An intermediate node must have the following characteristics: (1) freedom from flooding (broad cast storm problem), (2) limited power they can harvest or store, (3) ability to cope with node failures, (4) sustain during dynamic network topology, (5) predict communication failures, and (6) identify the duplicate packets.
A simple protocol design includes the header, payload and error detection fields in the wireless routing domain. In addition number of hop count is also considered for higher layer protocols like transmission control protocol (TCP) for setting the discard eligibility of the packet. A basic protocol design must be aware of the fundamental looping problem and sustain in node and link failure situation. Self-organized backbone network is defined as the backbone or sink nodes form a network, which should carry other node information during the routing process. The illustration of self-organized backbone nodes and networks in wireless domain is given in Fig. 1. From the figure, it is understood that the self-organized backbone node has high transmission power and are able to connect the backbone and normal nodes in one hop during the network failure conditions.
The processes like computing, table updations are account able factor for lifetime of the node in wireless networks. Each node must work with their own desire, conserve their resources and work in self-organized manner in MANET. For energy consuming, centralized approaches like cluster based routing not suitable because the whole network is based on the clusters as well as their coverage in mobile ad hoc networks. The two fundamental ideas: (1) selfish behavior of each node, (2) coverage of the node in ad hoc network motivated by us to develop this proposed work. To implement this approach all the nodes are permitted to work independently. Several algorithms [6], [7], [9] have been proposed to create backbone constructions. These algorithms deal with how to reduce the backbone in wireless networks and do not motivate the selfish behavior of the nodes. Modeling systems based on a selfish algorithm are developed to improve the performance of the ad hoc network, and do not guarantee the global solutions. The objective of the research work is more practical importance and has been underscored in [11], [12], [19] which provide excellent motivation to consider self-organization behavior of wireless networks.
Self-organized backbone set is a subset of connected dominating set. The CDS approaches follow the spanning tree or marking process. Maintaining the dominating set for a long duration is the main objective of the researchers in backbone concept. Self-organized backbones are created from the dominating set whenever the network needs to maintain the mobility. Complexity of the proposed approach is normally in logarithmic values only because the self-organized backbone involve in control transfer, where as CDS follows O (n2) complexity. The comparison of these two methods is given in Table 1.
The remaining of the paper is organized as follows: Section 2 describes the related work on the backbone construction with notations. Section 3, describes the formation of backbones. The selfish behavior of backbone node is given in Section 4 with tree structure. To illustrate the performance of the proposed network Section 5 is included with simulation results. Section 6, concludes the paper.
Section snippets
Related work
- •
The dominating set problem given in [2], [3] gives a survey on exact algorithms for NP-hard problems. It mainly focuses on exact algorithms for computing a minimum dominating set in undirected graphs. To find the smallest backbone is NP-hard, and show the time complexities of arbitrary graph is 1.93782n approximately. In our approach minimum connected dominating set (MCDS) i.e. self-organized backbones are obtained from the backbone set.
- •
Tseng et al. [4] and Lim and Kim [5] propose the solution
Formation of self-organized backbone network
The node formation in SOB-N is done in two ways: (1) marking scheme (2) spanning tree based. Construction stage of the network is based on self-organized backbones, which support alternate routing via one hop neighbors during link or node failure. Normal node movement, reconstruction of backbone and network updations are done in distributed manner. The next chapter shows the marking scheme based on self-organized backbone node construction with some theoretical proof.
Spanning tree based formation for self-organized backbone network
Let be an undirected graph for every bϵV let p(b) ⩽ 2 be a dual path between any two vertices. If p(b) = 2 the path follows dual tree otherwise shortest path between the nodes. Consider the problem of finding dual tree path (DT) between source and destination in unit disk graph. Dual tree formation is a sub graph H = (Vt, E), where Vt be the tree vertices and it may defined as Vt1, Vt2ϵV. Self-organized backbone formation is formulated using linear programming approach. Intermediate nodes
Performance metrics
The proposed approach self-organized backbone network (SOBN) is obtained from the backbones, and mainly concentrates on the activities of intermediate nodes. It supports power saving for individual nodes as well as the entire network and maintain the node mobility even if the last node fails in the network. The following parameters are used to analyze the performance of our work.
- •
Network size vs. SCDS: As the large number of nodes in a wireless network increases, scalability imposes difficulties
Conclusion and future work
In this paper, we presented possible architectural choices of self-organized backbones for mobility maintenance in ad hoc networks. The two construction methods marking scheme and dual tree based node constructions are analyzed in this work. Our proposal is comparable with existing routing protocol in all QoS aspects and compares (i) different network parameters (ii) scalability. It is the special idea since no additional cost and delay for selection of self-organized node from the backbone
Acknowledgements
The authors would like to acknowledge the Department of Science and Technology, India for sponsored the initial stage of this work to appear in the international conference under ITS scheme. Grant File No: SR/ITS/00377/2010-11.
S. Smys received the Masters in Network Engineering from Anna University, India. Currently he is an assistant professor in Karunya University, Coimbatore, India. His research interests are self-organization and topology control in mobile networks. He also a reviewer for International Journal of Wireless Personal Communication of Springer publication and publish more than fifteen research articles in Journals and IEEE conferences.
References (19)
- et al.
Flooding in wireless ad hoc networks
Comput Commun
(2001) A study of self-organization mechanisms in ad hoc and sensor networks
Comput Commun
(2008)- et al.
Modeling and computing throughput-capacity of wireless multihop networks
Comput Networks
(2008) - et al.
QoS routing with traffic distribution in mobile ad hoc networks
Comput Commun
(2009) - et al.
STAB-WIN- self-organized topology control ability backbone node in wireless networks
Wireless Personal Commun
(2010) - Gerhard J. Woeginger. Exact algorithms for NP-hard problems: a survey. Springer: LNCS 2570; 2003. p....
- Drori L, Peleg D. Faster exact solutions for some NP-hard problems. In: Proceedings of the seventh European symposium...
- et al.
The broadcast storm problem in a mobile ad hoc network
Wireless Networks
(2002) - et al.
Distributed algorithm for efficient construction and maintenance of connected k-hop dominating sets in mobile ad-hoc networks
IEEE Trans Mobile Comput
(2008)
Cited by (0)
S. Smys received the Masters in Network Engineering from Anna University, India. Currently he is an assistant professor in Karunya University, Coimbatore, India. His research interests are self-organization and topology control in mobile networks. He also a reviewer for International Journal of Wireless Personal Communication of Springer publication and publish more than fifteen research articles in Journals and IEEE conferences.
G. Josemin Bala received the Doctorate degree from the Anna University, India. Currently, she is a professor in Karunya University, Coimbatore, India. She is a member of IEEE. Her research interests include routing algorithms in mobile communication and complexity analysis in VLSI.
- ☆
Reviews processed and proposed for publication to Editor-in-Chief by Associate Editor Prof. Mehdi Shadaram.