Co-existence and spectrum sharing in White Spaces

One of the "hot" topics/question that came up at the IEEE Comm. Society meeting that was held in Washington DC on 19 September was the co-existence issues in TVWS (i.e. how spectrum should be shared among white space devices). The situation here is more complicated that it is the case for the licence-exempt access to, e.g., the ISM band because

-different access technologies may have to share TVWS spectrum among each other (e.g. 802.11.19. 802.11af, but also perhaps LTE, new standard for M2M etc)

-unlike the ISM bands where regulators have imposed low EIRP limits (100 mW in Europe, up to 4W in USA) to make possible efficient spatial sharing, in the case of TVWS we may see a mix of high and low-power use.

The main challenge is how to ensure fair (and also ideally efficient) sharing in TVWS between these heterogeneous users. The csma/ca-type listen-before-talk mechanism (these are sometimes called politness protocls or spectrum etiquettes) used in WiFi and Bluetoth are one option. This may work fine when sharing happens exclusively among  802.11-type systems but won't work if 4G technologies like LTE are also using these bands, since these technologies have very sophisticated mechanism for intra-system sharing but currently lack mechanism for sharing with other access technologies. Another issue with csma-ca type approaches is that because they are distributed a valuable portion of bandwidth is wasted for signalling. Also these mechanisms are ultimately aimed at best effort because they scarify QoS for fairness.

One possible short-tem solution to this problem is to use an additional layer in geolocation databases (GDB) that manage sharing between heterogeneous systems. This should be possible because GDB’s have access to information on location and type of devices. 

One potential problem, at least in the United States, with this approach is that the FCC does not seem to have mandated devices to report back to GDB provider the frequency and transmission power they are using. Another potential problem here is what happens to a database approach if sensing-only devices are also allowed in TVWS. More thoughts later on these and other issues of sharing.