After a very long delay, this year’s DySPAN was finally held two weeks ago in Bellevue, near Seattle. It was still pretty much focused on the TV White Space (TVWS). As far as I can remember, there was only one paper about the satellite band. The conference started with an one day tutorials on different subjects and followed by a three days mixture of keynotes, plenaries, poster/demos, and tech/policy sessions. During the conference, I came across a few fellow ex-QUASARians from BT, Ericsson and RWTH who were also presenting their results from the QUASAR project. A general impression that I felt during the conference is that, spectrum sharing is going to happen, but not as ‘the next big thing’ as some of us have hoped, nor would it be endorsed by any major operators anytime soon.
Out of the four tutorials on Tuesday. I attended the session on overview of policy and standardization in TVWS. The presenter, who was from Singapore, gave an panorama view of the progress in policy making regarding TVWS around the world: FCC is still the leading regulator who has published its regulation rule; Canadian has also proposed a similar framework to FCC, while OFCOM in UK and CEPT are still working together on the European level rule-making, which is expected to become public early next year. Another interesting development is the TVWS situation in Japan. There the TVWS under the typical definition does not exsit. Because there are many localized area broadcasting service for mobile TV on those TV WS channels. There is also a lack of drive for the industrial players to change this situation in Japan, since they are usually deeply involved with the TV broadcasting business already. On the industry standardization side, 802.11.af ‘super WiFi’ is still the most active group in the TVWS scope, and Singapore is having a national wide test run of a 802.11.af prototype system. The 802.19.1 coexistence in TVWS and the 802.15.4 are also emerging standards for TVWS systems, but they are less active than 802.11.af. However, if compared to the intense actions in 802.11. ac, the WiFi standard, even the development of 802.11.af would look like calm water. Upon reflection, this tutorial has already set the tone for the rest of the conference: stakeholder are skeptical due to uncertainty in regulation and aware of the cost for investment. TVWS will emerge and survive in the US but a global market is probably not going to develop.
The keynote speeches did not lighten up the prospect very much either. Two out of the five keynote speeches are from the policy side, and have discussed about which and how additional spectrum should be opened up for sharing, The President’s Council of Advisors on Science and Technology (PCAST) has been mentioned many times in both presentations. which is requesting for 1000MHz spectrum for shared use and they are leaning on the president to force the decision. However such passion in spectrum sharing does not seem to be shared by the industrial players, as all the other three presentations from Qualcomm, AT&T and Nokia have included little text about spectrum sharing in their slides, except for a few rather ambiguous comments. Apparently, the favorite solution for both Qualcomm and AT&T to meet the 1000 rimes higher capacity demand is still the installing of smaller cells. Whereas Qualcomm proposed a short cut by utilizing LTE broadcast and prefetching, AT&T is more concerned about the pathloss model and simulation tools that are suitable for small cell deployment. The presentation by Henry Tirri, CTO of Nokia, is quite different from the others, not only for its fancy graphics but also its long term visions. As far as I understood (which I am afraid is very limited), he first talked about his view on sensing in the wider sense. As we already have many different sensors in our mobiles, spectrum should and will be indexed by using sensor fusing just as the way how google map is indexing ‘where’ and facebook is indexing ‘who’. Therefore, he continued, it is not a problem how should we create the spectrum map, but rather how to use it once it is done. He further argues that any mechanism in spectrum sharing that involves manual input, like auction, negotiation, would be too inefficient to be applied to larger scales. They should be replaced by tools like algorithmic game theory to automate things. However, then the real question will arise, that is whether dynamic spectrum sharing is energy efficient enough for its gain in capacity. Because ultimately energy is the hard constraint in his belief. His intriguing and also confusing talk left many question marks in my mind but it is definitely one of the highlights of this conference, overshadowing the cruise ‘banquet’ that followed.
The technical and policy sessions did not seem to impress all of the audience, as I heard several comments like ‘interesting, but I did not learn anything new’ or more straightforwardly, ‘where is the revelation?’. This could possibly be explained by the delay of the conference and the fact that many of the papers presented were actually finished long time ago. Nevertheless, I still found quite a few interesting topics. One of them is the ‘gray space’ concept proposed by a PhD student from Oslo university. It basically suggest that we can use TV registration, which is also common in Sweden, to determine the location of the household that owns TV., and based on that information to better estimate the TVWS opportunity. There were also two interesting studies on TVWS in China. they claimed to have huge TVWS in urban china, which I find rather surprising. After discussing with the authors, I realize they actually means that there were too little TVWS in rural china due to the unfinished digital switch over process.
My own presentation was scheduled on the last day’s afternoon, but still around 30 people showed up and had a very interactive dissuasion. My talk was about the cumulative effect of interference from co-channel and adjacent channels on TVWS opportunity. The adjacent channel interference still appeared to be out of the radar in most of the research papers. If there is anyone who’d share my concern about adjacent channel interference, it is the TV broadcaster, and one of them was in the audience. She tried to use my result against unlicensed spectrum access in TVWS, but I assured her that it is not going to be an issue if TVWS devices are regulated properly under a secondary license. The session chair then posed a provocative question on whether TV receiver quality should be regulated as well. This time not the broadcaster but Reza Karimi from OFCOM came to the defense saying that it would be unfair to enforce TV receiver regulation just because we were trying to squeeze more interfering source next to the TVs.
After attending the last session, I rushed to the airport, and accidentally shared a cab with a senior from Interdigial. He was there supervising a demo by interdigital showing a wifi based discontinuous channel aggregation. On our way to the airport, he made a few comments that I believe could summarize many attendees’ feelings: TVWS as the test ground for spectrum sharing has made its most important contribution—the geolocation database; however the market for TVWS is struggling to develop, not because the higher layer system design or policies but due to the cost for the radio and the uncertain nature of secondary spectrum; spectrum sharing will emerge but is not going to be the prevailing solution. Upon his departure, he said,’ I was not being negative, but things are difficult, like all things are. Good luck with your study!’ Indeed, nothing is easy and there is no golden key that opens every door. Spectrum sharing might have worn out its halo, but it is still an interesting and meaningful subject for our research.