Designing systems for future obsolescence

The lifetime of government systems is far longer than the lifetime of the components that make up the systems. Obsolescence has become one of the normal design concerns for systems engineers. The primary factor in reducing the cost of redesign is an unambiguous, machine sensible systems description. Implementation independent tests can be generated from these descriptions, and they can be used in the design process as well. This can result in savings of up to 75% in the redesign of systems due to obsolete components (or the redesign of the components themselves). The design process consists of several phases: requirements capture, initial design, implementation, and verification. The verification step is by far the most costly, taking fifty to seventy percent of the system lifecycle resources. In a system design that takes obsolescence into account, the result of the initial design step is an implementation independent, unambiguous set of design requirements. These can be used as input to a tool, such as EDAptive Computing's VectorGen, to produce tests that automatically verify that the system meets its specification. In future redesigns due to obsolescence, the requirements and initial design steps are completely eliminated (because the implementation independent portion of the design is preserved) and the verification step is vastly reduced (because of the existence of implementation independent test vectors that verify that the device meets its specification). It is these features that result in the tremendous cost savings in future redesign efforts. This paper will detail the design process, describe the use of appropriate systems design methodologies and tools to produce the implementation independent design description given above, and document the potential cost savings.