Viewpoint

Design Reuse and Design Timescales

Design reuse provides considerable benefits, but care is required to minimize the initial slowing of the design process.

by Mark-Eric Jones

As ICs continue to follow Moore's Law and double in complexity every two years, the productivity of engineers is not keeping up with this geometric increase in gate count despite the advances in EDA tools and high-level methodologies. At the same time, an increasing proportion of all electronic design work involves IC design as levels of integration continue to rise.

Design reuse and the adoption of a "building block" approach to the IC design process offer hope to engineering managers struggling with design timescales in a market where reduction of product life-cycles is putting increasing time pressure on the design process. "Design reuse" is generally defined as a methodology whereby designs are created, documented, and maintained in an environment that allows blocks of the design to be easily and safely reused in future designs. Design reuse, however, is not a panacea that instantly guarantees shorter design timescales. Managers need to take care to ensure that this strategy doesn't initially slow down the design process. To this end, it is important to rapidly reach a "critical mass" of proven reusable building blocks that can be used in future designs.

That's not to say that design reuse won't produce long-term benefits; but in most organizations the advantages take some time to manifest themselves. In fact, initially, there will be a tendency for the methodology to slow down the design process rather than speed it up. The reason for this is that creating and documenting a design so that it can be reused adds a considerable overhead to the design process. It is only subsequent designs that will benefit from this investment, and the immediate effect on timescales is almost always negative; only after the completion and validation of this design can others benefit. It is important to be realistic about the practical effects of jumping on this particular bandwagon in order to minimize the negative factors and maximize the positive ones.

Design reuse results in the creation of a "fund" of proven design blocks that can be used in subsequent projects. In small and medium size companies, however, it may take some years for this fund to grow to a size where the benefits outweigh the overhead imposed by a good long-term reuse methodology. The fund of reusable designs will grow at a rate dependent on that at which new designs are completed, yet the size of the fund affects the rate at which these new designs can be created.

Engineers will recognize this as the equation for an exponential, with the growth of the fund at any time depending on its current size. Therefore it is clear that considerable advantages can be gained by not having to start the process at zero. The faster an initial fund can be acquired, the sooner design reuse will help rather than hinder the design process. Companies should look hard at the various ways in which they can "kick start" their fund of reusable designs, including modification of existing technology, joint venture projects, and technology licensing. These will be the key enablers that will bring forward the rewards promised by design reuse and overcome the initial overheads.

Mark-Eric Jones is chief executive officer and president of 3Soft Corp. (San Jose, CA).

integrated system design  April 1995