Abstract: If formal techniques are to be more widely accepted then they should evolve as current software engineering approaches evolve. Current techniques in the development of distributed systems use interface definition languages (IDLs) as a basis for the underlying communication and also as an abstraction tool. Object-oriented technologies [6] and the idea of engineering software through frameworks [5] are also widely accepted approaches in developing software. In this paper we show how the formal specification language SDL and associated tool support have been applied in the TOSCA1 project to engineer telecommunication services using these current techniques.
INTRODUCTION
Open distributed systems, i.e. (extendable) systems that interoperate to achieve some overall goal, represent a prime example of the kind of area where formal techniques could have a significant role to play. Whilst current technologies such as CORBA [3] have addressed many of the issues involved in developing distributed
systems, e.g. remoteness of components and their potential heterogeneity, such technologies fall short of being the final solution to building truly open distributed systems. CORBA allows system interconnectivity to be achieved, i.e. sub-systems will understand the messages that are sent to them, but this does not mean that they will interoperate correctly, i.e. work together to achieve some predefined goal. What they lack is behaviour.
Formal techniques offer a means whereby behavioural descriptions can be given both precisely and concisely. Unfortunately, most developers of open distributed systems rarely if ever apply formal techniques in the development of software. Why? Some of the more common reasons are:
• they are based on mathematical notations that are difficult to understand; • they produce models of systems that often bear no relation to the software itself
2;
• the models of systems produced are not usually re-used, whereas software – especially that based upon current practices such as object-oriented technology – is generally expected to be;
• they lack tool support for both developing and reasoning about the specifications.
In this paper we attempt to show through example, how SDL and associated tools as used in the TOSCA project address these issues. The rest of the paper is structured as follows. Section 2 provides an outline of the TOSCA project and the approach adopted to service creation together with an outline of the TINA architecture and the TINA object definition language (ODL). Section 3 provides an outline of the ODL/IDL mapping rules to SDL used in TOSCA. Section 4 presents the tool chain used in TOSCA. Section 5 provides an example of the development of a framework using this tool chain. Section 6 shows how the framework developed can be specialised to produce a model of a service. Finally, section 7 offers some conclusions and identifies areas of future work.
Dr Richard Sinnott
GMD Fokus Kaiserin-Augusta-Allee 31 Berlin, Germany sinnott@fokus.gmd.de
Mario Kolberg
Dept. Electronic and Electrical Engineering University of Strathclyde Glasgow, Scotland mkolberg@comms.eee.strath.ac.uk
No comments:
Post a Comment