While writing “The Ten Commandments for Effective Standards”, I stumbled across the University of New Hampshire’s InterOperability Lab. It’s quite an undertaking, and we’re all benefiting from their efforts. I wanted to learn more about what goes on at the InterOperability Lab and if it might pertain to EDA software. Today I had the pleasure of talking with Jon Beckwith and Bob Noseworthy, senior and chief engineers at the Lab. They gave me enough information to write another book, but for now I’ll just summarize our conversation.
The UNH InterOperability Lab was founded in 1988 and has grown into a major operation with around 120 people (100 are students) working in a 32,000 sq. ft. facility on some 26 programs in data networking and storage technology. The Lab is funded willfully by industry: companies pay fees (very reasonable amounts, IMHO, to cover expenses) and provide their latest products for inclusion in test beds.
Pretty much anything that has to do with networking can be tested for interoperability – i.e., compliance to standards – at the Lab. (Standards are the beginning of interoperability, but it’s conformance testing of products which implement a standard that uncovers limits, ambiguities, or missing parts in the standard.) The Lab delivers confidential reports to participating companies which tell them whether their products are robustly interoperable – or not. Product performance, price, features, and the like are not analyzed; that’s left up to the companies as they go about running their businesses.
For students working at the InterOperability Lab, it’s doubly beneficial. Not only are they paid for their apprenticeships, but they also gain real world experience that augments their studies. They work on a team which must produce real results for companies whose businesses depend on them. The students build professionalism and learn how to communicate because the reports they issue to companies could imply product cost and delay. Plus, having their hands on the newest products gives them an appreciation of how theories they learn are actually applied.
What has made the InterOperability Lab a success? Three things, certainly. The lab provides a high quality service, at an affordable price, from a trustworthy source (earned over time). Additionally, engineers from the InterOperability Lab do something special to ensure that their compliance testing is the best it can be. They actually participate in the development of standards. Their contributions during the creation of a standard center around the question, “Why standardize something if you can’t test it?” As the standard’s specification is constructed, the Lab guys ask, “How are you going to test this?” This additional dimension makes for a better standard and ultimately, better products that implement it.
Every once in a while, the concept of standards compliance testing for EDA products comes up. It’s been quite a few years since our industry attempted it for Verilog simulators at a third party interoperability lab. Would an interoperability lab for EDA fly today? It would be a complicated process to put together the countless design flows and tests that would cover the bulk of EDA tools and standards in the market today. Security would be a necessity before EDA companies and customers would feel comfortable contributing their tools and design test cases. It could be expensive, too, unless many companies were willing to chip in.
Another question should be asked, “Is independent compliance testing needed for EDA?” If EDA tools were seriously noncompliant and not interoperable, customers would likely scream (the way they did when they needed a standard for low-power design), wouldn’t they? Or perhaps the cost of nonconformance to standards in EDA isn’t significant enough to warrant an independent compliance lab. Or maybe EDA vendors are doing an acceptable job of preventing and correcting interoperability problems? (I can feel eyes rolling.) It could be that customers have learned to live with a measure of non-interoperability in favor of getting new products faster. It would be interesting to learn if an interoperability lab like that of the University of New Hampshire would bring a big benefit to EDA.
One thing I didn’t learn about the University of New Hampshire’s InterOperability Lab is why the “O” is capitalized. I’ll have to call Jon and Bob.