CYNTHIA JAYAPAL et. al.: ADAPTIVE SERVICE PROVISIONING FOR MOBILE AD HOC NETWORKS ADAPTIVE SERVICE PROVISIONING FOR MOBILE AD HOC NETWORKS Cynthia Jayapal 1 , Sumathi Vembu 2 and Arul Jothi Subramanium 3 1 Department of Information Technology, Kumaraguru College of Technology, India E-mail: cynthia_clement@rediffmail.com 2 Department of Electronics and Communication Engineering, Government College of Technology, India E-mail: sumi_gct2001@yahoo.co.in 3 Department of Computer Science and Engineering, Sriguru Institute of Technology, India E-mail: saruljothi@yahoo.com Abstract Providing efficient and scalable service provisioning in Mobile Ad Hoc Network (MANET) is a big research challenge. In adaptive service provisioning mechanism an adaptive election procedure is used to select a coordinator node. The role of a service coordinator is crucial in any distributed directory based service provisioning scheme. The existing coordinator election schemes use either the nodeID or a hash function to choose the coordinator. In these schemes, the leader changes are more frequent due to node mobility. We propose an adaptive scheme that makes use of an eligibility factor that is calculated based on the distance to the zone center, remaining battery power and average speed to elect a core node that change according to the network dynamics. We also retain the node with the second highest priority as a backup node. Our algorithm is compared with the existing solution by simulation and the result shows that the core node selected by us is more stable and hence reduces the number of handoffs. This in turn improves the service delivery performance by increasing the packet delivery ratio and decreasing the delay, the overhead and the forwarding cost. Keywords: Adaptive Algorithm, Leader Election, Mobile Ad Hoc Networks, Service Provisioning, Service Delivery. 1. INTRODUCTION Service provisioning in MANET is essential due to the rapid proliferation of mobile nodes and the need for delivery of services to these mobile nodes. Any user with a mobile device would like to avail any service at his convenience. A service is a software component executed in one or more nodes that reacts to service requests from several clients. A user would like to hear to music or download a file or would like to know the traffic details or entertainment information in a particular location. A service provider may provide one or more service for a short or long duration. It is the responsibility of the service provider to advertise its services. Different approaches are followed for service provisioning in the literature. In the advertisement based approach, a node providing the service advertises the service to all the nodes. The service requestor searches its cache to find a suitable service provider and forwards a request. The service provider delivers the service to the requestor. In the discovery based approach, there is no prior service advertisement. Service discovery is done by the service requestor, to identify the node providing the service and then it makes a request to the service provider, which delivers the service to the requestor [1]. In the directory based service discovery approach, a service provider registers its service with an agreed upon core node, either in a centralized or distributed fashion [2]. A service requestor contacts the core node to acquire the information regarding the service. The request is forwarded to the service provider and the service provider provides the service to the service requestor. A single service can be provided by different providers or a service request may need to be satisfied by several providers. All the available services and their providers must be coordinated. Centralized directory based schemes, [3] help in communication between the service provider and its clients, but is hard to scale and leads to bottlenecks. Local directory based scheme [4] search as request by distribution but is hard to scale for large networks. Delivering the appropriate service to the service requestor is often done by multicasting, as it is an efficient method to implement group communication in MANET. For multicasting various tree based, mesh based and hybrid algorithms have been proposed [5]. Latest literature suggests geographic position based routing as the efficient way of delivering service in MANETs, as it reduces the overhead in state maintenance of multicast trees [6]. Using hierarchical service provisioning in MANETs reduces the encoding and state maintenance overhead [7]. In group communication protocols, leader election is necessary when a group leader crashes or leaves the system. In context of MANETs, the leader election algorithm should eventually elect a unique leader and the elected leader, must be the most valued node among all the nodes in terms of performance related characteristics [5]. We propose an adaptive election scheme that chooses a leader to act as a service coordinator. 2. RELATED WORK 2.1 ROLE OF A SERVICE COORDINATOR The role of a service coordinator in service provisioning is vital. A service coordinator aids in service registration, discovery, coordination and delivery. Literatures [8]-[12], illustrate the role of coordinator in service provisioning. 2.1.1 Service Registration: A service coordinator can act as a service directory, where all the service providers register their service descriptions. It is the service provider’s responsibility to register and update the service information in the service directory. The service descriptions include information such as service providers identification, location, service identification, parameters describing the content of service and the life time of the service. 166