Network and Service Management
Current and previous projects in this area include
VESPER,
MANTRIP,
POLYMICS,
TEQUILA and
PAQMAN.
In recent years software mobile agents have been considered to be a promising line of inquiry in the area of network management. The research conducted here aims to present a clear direction for the practical exploitation of mobile agents for network management tasks. Three different case studies of network performance management have been examined in order to evaluate the effectiveness of the agent mobility strategy and autonomous behaviour applied within the specific context. The mobile agent benefits identified relate primarily to their easy support for programmability of network elements and the autonomous, self-configurable and fault tolerant agent operation. An important drawback however is that the advanced capabilities of modern mobile agent frameworks typically incur significant performance overheads, and these have been confirmed through a detailed performance evaluation comparing mobile agents with distributed object and mobile code approaches. To address this drawback, we propose network management solutions based on specially formulated execution environments that retain important mobile agent benefits while reducing performance overheads.
Policy-based management can guide the behaviour of a network or distributed system. This is achieved through high-level declarative directives that are dynamically introduced, checked for consistency, and refined and evaluated, resulting typically in a series of low-level actions. In this research we view policies as a means of extending the functionality of management systems dynamically, in conjunction with pre-existing "hard-wired" management logic. Inconsistencies in policy-based systems are quite likely since management logic is dynamically being added, changed and removed without the rigid development cycle of "hard-wired" long-term logic. Our work addresses the policy management aspects of architecture for managing IP DiffServ networks, focusing on the resource management part of the architecture. We have defined policies that drive the behaviour of an off-line network dimensioning component, as well as policies that can dynamically manage the resource reservations of the network devices based on actual network state and load utilisation. Finally, we have studied the hierarchical relation between these two types of policies, and we intend to specify conflict detection and resolution mechanisms specific to our problem domain.
The introduction of 3rd generation mobile communications networks (3G) will herald the widespread deployment of a new generation of mobile services (m-services). The distinct boundary between data and telecommunication networks, and similarly m-services, will become indistinguishable as mobile networks evolve towards a high bandwidth all-IP infrastructure. Future services are expected to be accessible through heterogeneous networks, with IP being deployed as the common network layer across multiple network domains. Furthermore, advanced service frameworks, such as the Virtual Home Environment (VHE) will allow users to access their services regardless of network used, terminal type and user location, particularly when roaming between such networks. These new developments require greater emphasis on service management in 3G networks and beyond. Of particular importance are new algorithms for service control, strategies and techniques for dynamic service adaptation, management frameworks for service configurability and service aggregation, and well-defined interfaces and platforms for service portability and accessibility during user roaming. This research will attempt to explore these issues fully with the aim of influencing the development of new management strategies and middleware in the future.
When advanced services are implemented for a mobile wireless context they require more complex management systems than the ones currently available. Therefore, with the increase of dynamic behaviour of networks such as ad-hoc and mobile networks, advanced services for next generation of communication systems need to be designed in a configurable way. In this research we are looking at services in which the individual components are realised with Mobile Agents (MAs). We present therefore a system which aims to manage those service constituents to meet issues such as:
- User requirements (including user profile and terminal profile);
- Service profile in terms of computational capability needed;
- Network QoS.
The main challenges of this research can be illustrated by summarising the work's objectives. Some of them have been addressed and some need further investigation:
- Investigate services and applications that are realised through the mobile agent paradigm;
- Study dynamic resource management algorithms for advanced adaptable services;
- Study the problem of evaluating such services and management algorithms and develop a suitable simulation methodology and environment for this purpose.
