Skip to main content

SecureNet - Network Security - Protection

Objective

The objective of the work is the systematic study and further development of methods to defend open computer networks against attack by malicious software, including viruses.
A project was set up to demonstrate the feasibility of the development of products for defending open computer networks against malicious software. The definition phase was divided into 5 work packages (WP) to be considered over 10 months. WP1 covered the collection and classification of threats of integrated broadband communication (IBC) environments namely, case studies of virus classification, modes of mobile viruses, detection capabilities and known weaknesses of IBC networks. WP2 covered the investigation of heuristic and formal algorithms for the detection, diagnosis and neutralisation, as well as selfcorrecting distributed computation (SDC) techniques for the provision of a trusted environment and the implications on IBC. WP3 consisted of the definition of an overall system concept, and WP4 covered the cost benefit analysis and selection of the most promising combinations of these techniques. WP5 covered the design and implementation specifications for a product (ie a set of tools) which incorporated the case selected in WP4. The project results indicated the necessity to attack the problem of detecting malicious software by monitoring system behaviour rather than by using the appearance of the malicious code itself. The wider applicability of the concepts in terms of analysing very large log audit files, for example, was also identified.
Technical Approach

The basic concept revolves around the co-operation of a neural net with an expert system and the use of self-correcting distributed computation (SDC) techniques.

The neural net component is responsible for the adaptive classification of events and the expert system uses the input from the neural net to perform a diagnosis and propose or perform neutralisation of any incurred damage. SDC techniques will be used to enhance immunity and to provide a minimum trusted environment.

During 1992, the project was in a definition phase whose success was intended to demonstrate the feasibility of the development of products for defending Open Computer Networks against malicious software.

The definition phase was divided into five work packages to be considered over 10 months. WP1 covered the collection and classification of threats of IBC environment namely, case studies of virus classification, modes of mobile viruses, detection capabilities and known weaknesses of IBC networks.

WP2 covered the investigation of heuristic and formal algorithms for the detection, diagnosis and neutralisation, as well as SDC techniques for the provision of a trusted environment and the implications on IBC.

WP3 consisted of the definition of an overall system concept, and WP4 covered the cost/benefit analysis and selection of the "most promising" combination of these techniques.

Finally, WP5 covered the design and implementation specifications for a product (i.e. a set of tools) which incorporated the case selected in WP4.

Key Issues

R&D is needed into methods of defence against specific forms of attack on software, notably "viruses" and "worms". This is because of the rising frequency of such attacks and the regrettably large number of people willing and able to devise and launch them. Viruses are important in view of the following:

Their extensive reproduction capability.
Detecting their origin is an undecidable problem.
They have the ability to completely alter (hence destroy) data.
Techniques for defence tend not to be preventive.

Expected Impact

The Project results indicate the necessity to attack the problem of detecting malicious software by monitoring system behaviour rather than by using the appearance of the malicious code itself. The wider applicability of the concepts in terms of analysing very large log/audit files, for example, was also identified.

Leaflet | Map data © OpenStreetMap contributors, Credit: EC-GISCO, © EuroGeographics for the administrative boundaries

Coordinator

Expertnet Ltd
Address
Dioharous Str 8
11528 Athens
Greece

Participants (6)

CCC SOFTWARE PROFESSIONALS
Finland
Dassault Automatismes et Télécommunications
France
Address
9 Rue Elsa Triolet Zone Industrielle Des Gatines
78373 Plaisir
Forschungsinstitut für anwendungsorientierte Wissensverarbeitung Ulm
Germany
Address
Helmholtzstraße 16
89081 Ulm
France Télécom
France
Address
38-40 Rue Du Général Leclerc
92131 Issy-les-moulineaux
Institute of Inform. Proc. Science
Finland
UNIVERSITY OF THE AEGEAN
Greece
Address
University Hill, Administration Bldg.