> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 1 Abstract— Mobile ad hoc networks (MANETs) form the underlying networking paradigm upon which pervasive and ubiquitous environments are founded. Combined with the growing need for mobility and flexibility in network infrastructures, the prominence and endurance of MANETs as a networking trend becomes evident. The main characteristics of MANETs include energy and bandwidth constraints, dynamic topologies and platform heterogeneity. Traditional network management principles and approaches are not applicable to MANETs, while on the other hand self-management principles are eminently suitable given the dynamic nature of MANETs. In this paper we propose, implement and evaluate a framework to enable self-management of MANETs built on an adaptive organizational model. We present the design of our proposed framework and we elaborate on the context-driven self- management cycle. We also present its evaluation through practical experiments using management case-studies. Index Terms—context awareness, mobile ad hoc networks, self organization I. INTRODUCTION HE proliferation of mobile ad hoc networking solutions experienced in the last few years and the high rates of user adoption of wireless technologies leads us to consider that there will be a paradigm shift from traditional, infrastructure-based networking towards wireless mobile, operator–free, infrastructure-less networking, with MANETs playing a key [4]. MANETs together with other emerging networking technologies, such as sensor networks, will constitute the foundations for future pervasive applications. The major strengths of this technology lie in the fact that it is easy to be deployed at a relatively low cost, while allowing for user creativity through the lack of central, authoritative management and control [7]. MANETs undoubtedly are not a solution for every networking problem of the emerging pervasive realm. Noteworthy drawbacks include their highly dynamic topology, since a node participating in a MANET is potentially mobile. These constant topological variations will This work was supported in part by the EU EMANICS Network of Excellence project (IST-026854). eventually lead to a continuous state of network instability, which in turn can deteriorate the performance of services and applications on these networks. It becomes evident that to fully benefit from the many potentials of ad hoc networking one needs to cater for a wide variety of requirements and provision for reliable, secure and efficient management [4], [7]. There is an obvious need for frameworks that can support the self-management of MANETs according to predefined goals or policies. We assert that such a highly dynamic environment can potentially benefit from context information that will drive its self-management, resulting in a degree of autonomy. The enabling technologies of autonomy and self- management also include policy-based management and programmability. Policy-based management provides the means to infer management decisions in a flexible and dynamic manner by matching contextual information to predefined rules. Programmability is also beneficial to enforce the required configuration changes on the system. This closed-loop adaptive management can thus lead to self- configuration, self-optimization, and hence a degree of autonomy. This paper addresses the design of a context-aware policy- based framework to achieve this, focusing mainly on its design and implementation aspects. The rest of the paper is organized as follows. After this brief introduction, related work is reviewed in section 2, while section 3 lays the foundations of our proposed organizational model. The design of our practical framework to enable self-management of MANETs is the focus of section 4. Implementation issues and related software metrics are presented in section 5 with indicative evaluation results presented in section 6. Finally, section 7 concludes the paper. II. RELATED WORK Autonomic computing refers to the self-managed operation of computing systems and networks, without the need for human administrators but with high-level objectives dictating the system’s functionality. The IBM autonomic computing blueprint [1] defines four distinct concepts behind autonomy, namely self-configuration, self-optimization, self-healing and self-protection [2]. The building block of all autonomic A Practical Framework to Enable the Self-Management of Mobile Ad-Hoc Networks Apostolos Malatras and George Pavlou Centre for Communications Systems Research, Department of Electronic Engineering, University of Surrey, UK a.malatras@surrey.ac.uk , g.pavlou@surrey.ac.uk T