The main aim of this project is to develop a distributed expert system (called EXPERNET) for the management of the National Network of Ukraine. EXPERNET will manage both the whole Ukrainian national network, as well as any of its sub-networks.It will consist of a number of local expert systems attached to related network management nodes, that belong to one of the three network levels: national, regional and district, closely reflecting the structure of the Ukrainian WANs. The design of the expert system will support both general and node-specific network management knowledge, in order to develop one expert system that can be easily adapted to the requirements of the specific node types. Furthermore, different interfaces between the expert system instances and the local WAN management software will be developed in order to facilitate the ease of information exchange between the two components of the local management software.The run-time aspects of EXPERNET include the exchange of information and resolution co-ordination between the physically and logically distributed local expert systems, based on global and local data and goals. The whole expert system is controlled either by data or goals by providing a possibility of selecting target modes dynamically for each node.

Importance

The development of an integrated national network infrastructure for Ukraine belongs to one of the most important tasks in accordance to the Informatization of Ukraine (1996-2000). The National Agency of Informatization of Ukraine will fund several projects during this period aiming to the integration of pre-existing networks into the new frame. However, the exploitation of any large WAN cannot be achieved without a user-friendly and intelligent network management software. Existing network management software cannot meet the requirements for the management of such large-scale networks. The successful implementation of this project can thus provide, apart from the products mentioned above, a pilot system for the intelligent network management of the whole Ukrainian country, that can be applied to other CCE/NIS or even EU countries, as well. Thus a potential marketing of the individual products and the system as a whole is foreseeable for the companies involved in the project.

Advantages

The advantages of EXPERNET over a centralised system are:

• The encapsulation of knowledge about the sub-networks at the target management nodes.
• The structure homomorphism between the knowledge base and the database of the conventional network management software.
• The parallelization (through distribution) of the expert system’s tasks.

Functionalities

The system will provide functionalities on all levels of network management, like:

• Support for administrative and operative WAN management.
• Consulting and educational orientation.
• Real-time network management.
• Control action for network nodes and devices.

System Components

In order to achieve such a functionality certain subsystems will also be developed.These include:

• Cs-Prolog II (CSP2) for UNIX workstation network in order to capture the distributed and real-time requirements of network management with a high-level, declarative programming language (ML Consultanting & Computing Ltd.).
• A knowledge model, that will be supported by a computational platform, able to represent and reason about the real-time, asynchronous and constantly changing knowledge of network management (Universidad Politecnica de Madrid).
• An active Knowledge Base Management System (Device) able to represent and reason about the real-time and event-driven network environment with a user-friendly shell that combines the power of the underlying KBS with the advanced programming environment of CSP2 (Aristotle University of Thessaloniki & ML Consultanting & Computing Ltd).
• A knowledge base able to represent both general and specific experience knowledge about network management and to explain the network staff about the optimal network operation (Aristotle University of Thessaloniki).
• Adaption of existing network moritoring tools, HNMS and Big Brother, as to serve as components of the final system (Technososoft & Glushkov Institute of Cybernetics).

Partners

• Intelligent Systems group, Dept. of Informatics, Aristotle University of Thessaloniki, Greece (coordinator)
  • I. Vlahavas
  • Ilias Sakellariou
• S.A., Thessaloniki, Greece
  • John Damianos
  • Panagiotis Grilas
• Dept. of Artificial Intelligence, Technical University of Madrid, Spain
  • Martin Molina
  • Sascha Ossowski
• ML Consultanting & Computing Ltd., Budapest, Hungary
  • Ivan Futo
  • Janos Szeredi
• International Software Technology Research Center TECHNOSOFT Kiev, Ukraine
  • Igor Velbitskiy
  • Sergey Yershov
  • Igor Netesin
• Glushkov Institute of Cybernetics, Ukrainian Academy of Sciences, Kiev, Ukraine
  • Vladimir Skurikhin
  • Anatoly Docoshenko