2012 2nd IEEE International Conference on Parallel, Distributed and Grid Computing
Cost Optimized K-Fault Tolerant Channel Assignment in Wireless Network Abdul Quaiyum Ansari Department of Electrical Engg Jamia Millia Islamia New Delhi, India
[email protected]
Premchand Saxena School of Computer & Systems Sciences Jawaharlal Nehru University New Delhi, India
[email protected] Kayel Datta Gupta Dept of Computer Science & Engg. Maharaja Surajmal Inst.of Tech. New Delhi, India
[email protected] to [7] and is called Link Deficit algorithm. The authors [8] proposed and analyzed new irregular Fault-tolerant multistage interconnection network which is dynamically re-routable and provide mUltiple paths of varying lengths between source and destination pairs. The paper [9] has considered the topological optimization of communication networks subject to reliability constraints. In [ 10] the author presents a decomposition approach to the problem where the total network cost is minimized considering the reliability constraint. The work [ 1 1] presents a systematic method for nwnbering the nodes assumed to be placed in a straight line with equal space between one another and the connectivity that is sought is even. The cost of a link has been asswned to be proportional to its length and claims to generate the cheapest starting network. The paper [ 12] proposes a design technique to tackle the fault tolerant problem of Mobile Adhoc Networks. This paper concentrates on two constraints that is communication cost and fault tolerance of a wireless network. For simplicity it has been assumed; that all the radio links have the same reliability which is between 0 and 1. In this paper, a technique for designing a wireless network with overall mlllllllum communication time has been developed. Through simulation it has been shown that the proposed algorithm can efficiently fllld out an optimal solution for the above problem. The paper is organized as follows. The preliminaries necessary for this work is presented in Section 2. Section 3 defines the problem statement. Section 4 presents the proposed algorithm. The outcome of the proposed algorithm OWN is illustrated with an example of synthetic data in section 5. Result analysis of the proposed technique is done in Section 6. Section 6 discusses the result analysis. Section 7 concludes the paper.
Abstract -The paper addresses the problem of disruption of services of any wireless network due to failure of a radio channel. This can be achieved by providing more than one disj oint route between any pair of node. Though higher redundancy ensures better reliability of the network but it also increases the network cost. In this paper, we propose a technique to design a cost optimal wireless network layout which incorporates some real life constraints of cost and reliability. A wireless network layout is only accepted when it can endure at least K radio channel failure and the total cost of the layout is minimum. In this technique n X n matrix is used to represent the cost of communication between the respective nodes which eventually is the communication time between a given pair of nodes. The technique is verified with a set of synthetic data. Keywords: Fault tolerance; Communication cost; Radio channel;
I.
INTRODUCTION
The notion of wireless mesh network is one in which each node is connected to all adjacent nodes rather than to a few switching points. The advantage of a mesh configuration is that it can survive failure of nodes, links or combinations of nodes and links. The nodes in the wireless network are asswned to be static and they communicate through multi-hop ad hoc radio channels by using intermediate nodes to relay the message. Thus, each node in the wireless network also acts as a router, and forwards data packets to other nodes. We consider that each wireless node has an omni directional antenna. A failure in a radio channel (due to synchronization) may disrupt a communication. Hence an important design constraint for any reliable wireless network is that the network must remain connected even on failure of one or more channels. The design of network layout is subjected to the cost constraints as well as the reliability requirements to cope with the fault that may occur in the network. A nwnber of approaches to topological network optimization have been developed [ 1], [2], [3], [4], [5]. The work [6] develops a channel selection algorithm and integrates the same into the distributed acquisition algorithm. A popular heuristic for generating a minimal cost starting network is due
978-1-4673-2925-5/12/$31.00 ©2012 IEEE
II.
PRELIMlNARIES
A. Some Relevant Definitions A channel failure is an abrupt collapse of a channel due to some unstable condition which leads to interruption in communication.
18
2012 2nd IEEE International Conference on Parallel, Distributed and Grid Computing
A network topology is K-FT iff every pair of nodes is reachable for any K link failure in the network or there are K+ 1 disjoint paths between any pair of nodes. A graph G is connected iff every node in G is reachable from all other nodes. A graph is K-FT iff all the graphs, which have K less link than graph G are connected. The degree of a node is defmed as the number of links/edges incident on it.PROBLEM STATEMENT Given n no. of nodes and Cost matrix Costjj for all i, j then find a K-FT network topology such that the total communication cost is minimized. Assumptions 1. The geographical location of the nodes is known a priori and the position of the nodes are stationary. 2. The reliability of the nodes are 1. 3. The communication cost between each pair is given. 4. The channels are bidirectional. Costjj = Costjj for all i,j En && Costjj = w. 5. All radio channels have equal bandwidth. 6. All radio channels are mutually non-interfering. 7. The network layout uses a unique spectrum between any given pair of node. In this paper the term channel and link has been used mutually interchangeably. The problem is to find an optimal network layout for the given set of nodes such that the communication time is minimized, subjected to the condition that the layout can endure K number of channel failures. Given n and Costij for all i, j the problem is to assign radio channels to the node such that network topology is K-FT and the total communication cost is minimized. III.
A.
System variables
The following variables have been used in the proposed algorithm n: number of nodes K: Maximum number of radio channel failures permissible without performance hindrance I::SK::Sn-2 Ii: Initila node ordering (random in nature) Costjj : Cost of communication between node and node j where i'J E n && j:;ti nodej: it node Link_Sumj: Sum of node i's communication cost ¥ node j where i,jE n && j:;ti Min_Costj: The minimum cost of a link assigned to node i where iEn Tot Cost: Cost of the final network Cointj = number of radio channels currently allocated to node I.
L = list of assigned links B.
Algorithm
OWN
n = number of nodes Costij = cost of a link between node i and j for every pair of i and j. K= no. of channel failures that can be endured by the network Output: K-FT network layout with optimized cost. Step 1: //Initialization Tot Cost=O L =0 for i = 1: n { Link_Sumj =0 // Initially link sum is 0 Min_Costj = W // minimum cost is assigned to w Countj = 0 // no link is assigned to any node yet } Step 2: for i = 1: n // Finding the sum of links and minimum link cost for each node for j= l:n if j:;ti { Link _Sumj + = Costjj // the cost of radio channel for each node is calculated If Min_Costj > Costij // the minimum cost of any radio channel for a given node is determined Min_Costj = Costjj } Step 3: Link _Sumj is sorted in descending order using Quick Sort. Step 4: for x=l:n { // All nodes are numbered from 1 to n with highest //Link _Sumj value to lowest. Ij