PAWS; radios and databases speaking the same language (but there can be many dialects)

2nd October 2015

The Ofcom pilot was a tremendous opportunity for us to test the technology and scenarios for TV white spaces (TVWS). We have used the technology to provide backhaul connectivity for the Oxford Flood Network; designing, deploying, monitoring, and maintaining the network allowed us to gain in-depth technical knowledge and has provided us with a number of ideas for the future. We also investigated with other partners (King’s College London, Eurecom, and JRC) how the proposed UK regulatory framework would handle mobile nodes. Interestingly, similar questions have been researched by Prof. Pawelczak from TU-Delft, and he used our pilot White Spaces Database (WSDB) for one of his academic publications; if you are curious to know more about this topic, check his publications list.

Nominet is currently going through a qualification phase with Ofcom so that our geo-location database will be ready to operate in the UK at the beginning of 2016.

While we are preparing, I thought it would be interesting to tell a bit more about how the geo-location databases and TVWS radios talk to each other

As I mentioned in a previous post, devices wishing to operate in the TVWS need to first contact a geo-location database operating in that area to obtain the operational parameters so that they do not to cause interference with digital TV transmissions and other licensed devices such as wireless microphones. In practice, the more databases that devices can talk to, the better. At the same time, geo-location databases want to make sure that they can interface with as many devices as possible without having to provide bespoke APIs for each device manufacturer. This means that radios and geo-location databases need to speak the same language. In other words, there is a need for a standard communication protocol.

Fortunately there is such a standard, the unimaginatively named Protocol to Access White-Space Databases (or simply PAWS), published by IETF as RFC 7545. Although developed with TVWS communication in mind, it could be actually used for other portions of the spectrum as long as the general communication paradigm (based on geo-location database, master devices and slave devices) still applies.

PAWS specifies how a master device obtains a schedule of available spectrum at its location; it also takes into consideration the security necessary to ensure the accuracy, privacy, and confidentiality of the device’s location. Being an IETF standard, the protocol assumes that the master device and the WSDB are connected to the public Internet (or private IP networks).

PAWS recognises that different regulatory domains may impose particular requirements. This means that the actual implementation of the protocol might change among different countries. In other words, for the same underlying language (PAWS) every regulatory area might have its own dialect. While writing this post I was reminded of an interview with Hugh Laurie and Ellen DeGeneres that illustrates the challenges of different communicating using different dialects:

Imagine how difficult it would be if devices and databases spoke with different accents, or dialects. Information could be interpreted in slightly different ways, creating lots of confusion. In creating a flexible PAWS the only way to avoid recreating a TVWS Tower of Babel is to make sure that device manufacturers and WSDB providers operating in the same regulatory region agree to use the same PAWS dialect.

The Babel tower

Figure 1: The Babel tower (source)

Fortunately Nominet and other companies are working to standardise the dialect that TVWS radios and geo-location databases will use to communicate in the UK. Many questions – such as “What is the minimum set of messages that radios need to support to operate within the local regulatory framework?”, “How does the location uncertainty units used by local regulations translate to the PAWS protocol?” – are being discussed and common approaches agreed.

The companies working to launch the geo-location databases in the UK have agreed on a document depicting the supported message scenarios, which has been shared with Ofcom. We are now starting to map these messages to the PAWS protocol and, once agreed on the details, we will modify the WSDBs to be able to speak the same PAWS dialect. Once these documents are finished, we will seek input from the
device manufacturer, and we believe this process will facilitate the growth of the TVWS market in the UK and hopefully in Europe.