Listening to the talks at the SDR Europe today in Brussels on developments of Software Defined Radios makes me somewhat skeptical. Reprogramming the radios to be able to quickly change the radio access waveforms and protocol sounded like a great idea 10 years ago when there were many competing “3G” air interfaces for mobile communication. As we now see a clear convergence of massproduced equipment to two interfaces, one LTE-typ interface for wide area systems and one that will somehow be containing the numbers “802” and “11”, for short range communication, there seems to be little space for the use of SDR:s. The industry clearly went the other way and smartphones today contain 6-7 different (non-programmable) radios – mainly driven by the fact that the various standards for which they were designed used different frequency bands. SDR:s do not solve the spectrum issue – it is still very hard to build a radio (RF front ends), programmable or not, that can operate over an octave of frequency with reasonable performance.
Prototyping, small volumes, fine! I can even understand the reasoning in the military domain where you need backward compatibility with legacy equipment, “horizontal compatibility” with other systems of similar character in the same frequency band (e.g. used by other armies you would like to cooperate with). However, this is not quite what the industry sells. Many seem to misunderstand the basic problems of “forward”compatibility. Of course you would like to buy a “future proof” radio that can in the future be reprogrammed with the latest functionality and a more capable air interface. But this has never happened and will never happen simply due to the fact that high performance, wide band SDR:s already operate at the limits of their performance. Its the hardware that is the limitation, not the ingenuity of our signal processing engineers. This is very different compared to microcontrollers in a car or dishwasher were the computational platforms are several orders of magnitude more capable than the actual needs. The consequence is that the day the next improved air interface comes along, your SDR hardware is not capable enough. This is exactly why, after a few years, you cannot upgrade your once top-of-the-line PC to the latest Windows software – it has too poor performance and too little memory. So why is ETSI spending all this time and effort to standardize a civilian “reconfigurable mobile device” architecture? What was the use case? Does it still exist?