Estimating Reliability of Mobile Agent System for Mobile Ad hoc Networks Chandreyee Chowdhury Sarmistha Neogy Jadavpur University Kolkata, India email: chandreyee.chowdhury@gmail.com email: sarmisthaneogy@gmail.com Abstract- A mobile agent is a particular type of agent with the ability to migrate from one host to another where it can resume its execution. Mobile agents can be used in applications for wireless network to save bandwidth and time. In this paper we consider dependability aspect, especially reliability issues that need to be addressed before the mobile agents can be used for a broad range of commercial applications in mobile adhoc network (MANET). The MANET introduces new aspects to dependability affecting reliability of the services of the mobile agent based system. In this paper, we propose an algorithm for estimating the task route reliability of mobile agent systems (MAS), which are based on the conditions of the underlying wireless network. The agents can select their route dynamically. The complexity of mobile agent network systems makes it hard to obtain the route reliability of the MAS theoretically; instead we estimate it using Monte Carlo simulation. In this paper, we assume that the MAS consist of a number of independent mobile agents operating simultaneously. The paper estimates the reliability of mobile agent system considering different failures of the underlying network. The results achieved demonstrate the robustness of the proposed algorithm. This paper demonstrates a reliability estimation model for MAS in MANET and shows that reliability is almost independent on the no. of agents (for higher no. of agents) or no. of mobile nodes but highly dependent on the condition of the links especially on the connectivity of the graph. Keywords- Mobile agent; dependability; reliability; Monte Carlo. I. INTRODUCTION A mobile agent is a combination of software program and data which migrates from a site to another site to perform tasks assigned by a user according to a static or dynamic route. It can be viewed as a distributed abstraction layer that provides the concepts and mechanisms for mobility and communication [1]. An agent consists of three components: the program which implements it, the execution state of the program and the data. A mobile node may migrate in two ways namely weak migration and strong migration. Weak migration occurs when only the code of the agent migrates to its destination, a strong migration occurs when the mobile agent carries out its migrations between different hosts while conserving its data, state and code. The platform is the environment of execution. The platform makes it possible to create mobile agents; it offers the necessary elements required by them to perform their tasks such as execution, migration towards other platforms and so on. Typical benefits of using mobile agents include • Bandwidth conservation: sending a complex query to the database server for processing. • Reduced latency: a lightweight server can move closer to its clients • Load balancing: loads may move from one machine to the other within a network etc. The route of the mobile agent can be decided by its owner or a mobile agent can decide its next hop destination on the fly. Here, we assume the underlying network to be a Mobile Ad Hoc Network (MANET) that typically undergoes constant topology changes, which disrupt the flow of information over the existing paths. Mobile agents are nowadays used in MANETs for various purposes like service discovery [2], network discovery [3], automatic network reconfiguration etc. Dependability of any computing system may be defined as the trustworthiness of the system which allows reliance to be justifiably placed on the service it delivers [4]. It is an integrative concept that encompasses attributes like availability (readiness of usage) and reliability (continuity of correct service) [4]. In MANET, like in any other mobile distributed system, mobile nodes access information through wireless data communication at any time and everywhere (motion and location independence) [5]. Therefore, this environment itself introduces new features and aspects to dependability, affecting both availability (readiness for usage) and reliability (continuity of correct service) of the services of distributed systems. Hence the reliability of underlying network becomes a factor that may affect the performance, availability, and strategy of mobile agent systems [6]. In this paper, we define a Mobile Agent-based System (MAS) to be a system consisting of a number of independent agents, each of which accomplishes an independent task. We study the reliability of a mobile agent system with respect to the network status and its conditions. We propose a randomized agent planning strategy where an agent selects a destination randomly and the routes are also updated dynamically as agents roam the network in order to reflect node mobility. Our work in reliability estimation is presented in details in the following section (II). The simulation results of our reliability model are summarized in section III. Some contemporary work in this area is discussed in section IV. Finally section V concludes with an indication of our future endeavor in this area. 2010 Third International Conference on Dependability 978-0-7695-4090-0/10 $26.00 © 2010 IEEE DOI 10.1109/DEPEND.2010.15 45