Autonomic Self-Healing for MANETs J. Chaudhry 1 , Y. Lee 2 , K. Pence 3 , J. Sztipanovits 2 1 Università degli Studi di Trento, TN, Italy 2 ISIS, Vanderbilt University, Nashville, TN, USA 3 EECS-EngM, Vanderbilt University, Nashville, TN, USA Abstract – Self-healing systems are considered as cognation-enabled sub form of fault tolerance system. But our experiments that we report in this paper show that self- healing systems can be used for performance optimization, configuration management, access control management and many other functions. The exponential complexity that results from interaction between autonomic systems and users (software and human users) has hindered the deployment and user of intelligent systems for some time. We show that if exceptional complexity is converted into self- growing knowledge, (policies in our case), can make up for the initial development cost of building an intelligent system. In this paper, we propose that AHSEN (Autonomic Healing- based Self management Engine) clearly demarcates the logical ambiguities in contemporary designs and shows its performance through empirical results obtained through experiments. Keywords: Automatic Self-Healing, MANETs, Congestion Control 1 Introduction As the complexity and size of networks increase so do the costs of network management [1]. The preemptive measures have done little to cut down on network management cost. Hybrid networks cater with high levels of Quality of Service (QoS), scalability, and dynamic service delivery requirements. The amplified utilization of hybrid networks i.e. ubiquitous-Zone based (u-Zone) networks has raised the importance of human resources, down-time, and user training costs [10]. The u-Zone networks are the fusion of the cluster of hybrid Mobile Ad-hoc NETworks (MANETs) and high speed mesh network backbones. They provide robust wireless connectivity to heterogeneous wireless devices and take less setup time. The clusters of hybrid networks feature heterogeneity, mobility, dynamic topologies, limited physical security, and limited survivability [2] and the mesh networks provide the high speed feedback to the connected clusters. The applications of MANETs vary in a great range from disaster and emergency response, to entertainment and internet connectivity to mobile users. Autonomic Computing provides a cheaper solution for robust network management in u-Zone networks in the form of self-management. Self Management is a tool through which performance of the computer systems can be optimized without human user intervention. In [24] Turing et. al. suggests that autonomic systems have exponential complexity which can hamper the appropriate problem marking and also raises the software cost. So it is critical to provide incremental, low cost and time efficient solutions along with minimizing the maintenance cost of the software. The u-Zone networks contain a highly vast variety of devices connected to them. It is not apposite to address the problems of each category of devices individually. We need to have some general solutions that could entertain a certain set of devices. Moreover, the probability of a management solution made for one type of client would be appropriate for another client is very low. The authors in [17] target the self management in hybrid environment through a ‘divide and conquer’ approach by using component-based programming. They propose to rapidly divide the problem into sub-domain and each domain is then assigned ‘sub solutions’. The amalgamation of all ‘sub solutions’ gives the final management solution to the client. Several network management solutions proposed in [4] [5] [6] [7] are confined strictly to their respective domains i.e. either mesh network or MANETs. A self-management architecture is proposed in [3] for u-zone networks. We have identified the following questions that are still to be answered since there is only a limited amount of published work on the topic: 1. If self-healing is one of the FCAPS functions (Fault, Configuration, Accounting/Administration, Performance, Security) than what is the physical location of self- healing functions whether it should reside on the gateway or at the client end? 2. How does the control, information etc flow from one function to another? Especially how do Self-healing functions interact with the other functions? 3. What are the calling signatures of self-healing functions? If Self-healing functions are fault-removing functions, than what are the functions of Fault Management functions? 4. Are these sub-functions functionally independent? If yes, then there is an evidence of lot of redundancy and if not then how self-healing can be thought of an independent