SPECIFICATION DRIVEN BEHAVIORAL DESIGN OF COMPLEX SYSTEMS

Article type: Research Article

Authors: Tanir, Oryal

Affiliations: Director - Business Engineering and Simulation, Bell Canada Montreal, QC, Canada

Abstract: Traditionally in the design automation field, simulation has been applied to explore low-level design details. However, a focus upon design specification driven simulation environments has been missing. Such tools can conceptually capture an initial design specification and allow design exploration at a highly abstract level with a simulation-based environment prior to synthesis. At this design stage, software and hardware elements can be indistinguishable and represent a good opportunity for addressing system level concerns such as partitioning, co-design and architectural trade-offs. With the advent of more complex systems, the ability to model and verify properties of various alternative designs is mandatory to produce cost effective and sound systems. Specification driven design implies a need to support architectural design and rapid prototyping systems (RPS) within a design flow. The system design activity is generally the starting point within the design phase of a product life-cycle - which involves the design capture of specifications into an executable model. Hence language requirements at this stage encompass modeling capabilities. The design and modeling of the conceptual system at this abstract level implies that the design environment must support concepts such as generic model reuse, component and structural reuse, intelligent library management, and hierarchical design. After a suitable model is defined, the language must provide support for experimentation and analysis. These activities are crucial for a designer to explore a given design space, make appropriate trade-offs and partition the design to different hardware/software configurations. Such activities can be supported through design simulators and formal methods. This paper examines the benefits of applying a specification driven approach and presents a framework for environments that can support the related design activities. The Design Analysis and Simulation Environment (DASE) based upon this framework has been successfully implemented through a joint initiative between Bell Canada and McGill University.

Journal: Journal of Integrated Design & Process Science, vol. 5, no. 2, pp. 1-11, 2001