COTS on the wrong road
There is another aspect to your article entitled "Could the COTS movement be heading in the wrong direction?" that appeared in the June issue of Military & Aerospace Electronics.
To The Editor:
There is another aspect to your article entitled "Could the COTS movement be heading in the wrong direction?" that appeared in the June issue of Military & Aerospace Electronics. Several people here say that COTS has degenerated until it now stands for "Currently off the shelf" electronics. My first experience with that was on a U.S. Navy design several years ago.
The previous design of the product in question used a custom Z80 microprocessor-based system. For the new design we needed a microprocessor design and wanted to go COTS. There are many 'one board' computer systems available, but none of them are form, fit, and function interchangeable.
We decided to go with the PC-104 form factor where we could find multiple vendors for each type of board we needed. We put together a five-board stack and started the software design, packaging design, and the design of the other components.
The good news was that by the time we finished the design new PC-104 boards were on the market that reduced our system board count from five to four. The bad news was that only one of the four was one of the original boards selected and the software had to be tweaked to accommodate the new boards.
The concern now is how often will the software need to be tweaked to accommodate replacement for boards that are no longer available? For military systems with long projected life spans, frequent changes of PC Boards and software are major configuration management problems. The cost of the configuration management paperwork can exceed any cost savings furnished by newer processor boards. The military is never able to purchase enough new widget version 2 to outfit the entire service and install them in one year.
Our choice seems to be to have a design stable long enough to produce and install it over several years, put maintenance training in place, and then use the system for a few years or to make the system disposable, have no maintenance training, and every few ships will then have a different model of the system. At that point how do you handle operational training on all of the different versions of the system?
Design is the smallest part of the effort to create a system. Production of the systems, installation of the systems, construction of the maintenance and operational training infrastructures, and then the service life of the equipment all span a longer time. The design needs to remain stable long enough for the equipment to still have some useful field life.
This is becoming increasingly difficult as the throwaway commercial electronics market drives parts availability. Yes, COTS really now means currently off the shelf.
Johns Hopkins University Applied Physics Laboratory