FINITE (Feature Interaction in Telecommunications)

This project is being undertaken from September 1993 to May 1997. The goal is a architectural method for modelling and analysing telecommunications architectures with respect to the feature interaction problem. The project is being undertaken by Nikolaos Kosmas under the supervision of Ken Turner.

Introduction

The areas of interest that are directly related to this work have a wide spread. For even just telecommunication architectures and models, influences come from the Intelligent Network (IN), Open Systems Interconnection (OSI), Open Distributed Processing (ODP), Broadband Communications Networks, and Communications Management Architectures. Because Europe is becoming more active in the area of communications, it is appropriate to consider European projects and approaches such as the Telecommunications Networking Information Architecture (TINA) and the Race Open Services Architecture (ROSA). These architectural frameworks are combined with analysis and design techniques (object-oriented techniques, formal methods) and other generic frameworks (Open Distributed Processing) to offer perspectives on analysing telecommunication architectures.

The focus of the research is Feature Interaction (FI), with many starting points towards a solution to the problem. The goal of this research is therefore analysis of telecommunications architectures with respect to the feature interaction problem. The analysis and/or design of telecommunication systems requires huge investments of manpower and resources. Therefore, they are not good targets for research at least on an individual basis. Consequently, work is focused on directions such that expertise and other research in the Department can be utilised.

Current Work

• Formal methods LOTOS, SDL and Z have been investigated for their suitability. Tools have been used for the production of specifications and prototypes. Some preliminary efforts to formalise a POTS service were successful as first examples, but not sufficient for immediate practical use on the Feature Interaction problem. It was shown that a more generic approach making use of formalised concepts from telecommunications and other areas is necessary.
• Object-Oriented (OO) Techniques Work has been based on research previously done in the Department by other colleagues. It was considered important to explore architectural development with respect to formal methods and object-oriented techniques. For the latter, previous experience on real-world examples (mainly by BT) has been examined with special interest. Tools like OMT and Object Tool are currently being used to produce object models. Other relevant frameworks based on the emerging field of patterns in object orientation are also being considered.
• IN architecture The concept and structure of the IN architecture has been a point of interest and so has been covered in detail. Models based on conventional software engineering techniques have been created, and weaknesses (in terms of principles) of the IN structure and architecture have been identified. Concepts from the IN approach are being integrated within an architectural framework including Formal Methods, OO and ODP. Thus, a mapping between these areas is a vital aspect of the work.
• Open Distributed Processing (ODP) Due to the relationship of the work within ODP to telecommunications, it was considered crucial to investigate the applicability of the principles and methodologies of ODP in to the Feature Interaction problem. The ODP framework and its modelling techniques have been evaluated.
• Telecommunication concepts Architectural approaches and general information about telecommunications (ATM switching, management of telecommunications networks) have been analysed and embodied in a general framework. The main aim was to develop a sound telecommunication context that covers all possible considerations. This has included identifying the major forces that are responsible for the tremendous evolution of telecommunications (e.g. deregulation, privatisation) and the future plans of the majority of providers and vendors concerning the wider market. This was considered desirable for two reasons: to avoid research that does not correspond to reality, and because of the influence of market factors on technical aspects of telecommunications (e.g. the Customer Premises Equipment (CPE) market and the user role).
• Feature Interaction (FI) General approaches to the FI problem have been investigated. Since the FI problem is too large to address as a whole, a classification of Feature Interactions has been made on the basis of a number of criteria.

Future Work

Future work will concentrate on:

• Architectural Development architecture in the conventional software engineering cycle; architecture as a driving force; definition of the new architecture; elements and semantics of architecture; relation between architecture and engineering phases; architectural principles and design; architecture in a service engineering context
• Network Architecture evaluation of the IN architecture; definition of a generic telecommunication architecture; architectural and conceptual models; high-level concept hierarchy; architectural components and combinators; interpretation and projection using the ODP viewpoints; formal models; architectural semantics
• Service Creation Process definition of service, service feature, service creation environment; generic services; encapsulation of call-control features into service objects; deployment (activation/deactivation) of services; classification and specification of services; service interfacing; structural decomposition of services; redefinition of a service engineering discipline
• Feature Interaction definition of problem domain; classification and categorisation of feature interactions; analysis of specification level interactions; influence of feature interactions on the underlying model; recognition of pattern behaviour in interactions using service specifications; development of detection heuristics; resolution issues
• Software Tools use of Rigorous Object-Oriented Analysis (ROOA), Specification using an Object-Oriented LOTOS-Based Visual Language (SOLVe), LOV-OMT, GEODE, LOTOSPHERE Integrated Tool Environment (LITE)

Up one level to Ken Turner - Research Projects