Applications for white space spectrum?

The search for what to do with white space spectrum is on.

Cambridge Consultants, a radio consulting company in UK, identified three promising applications: in-home video distribution, municipal wireless, and rural broadband. Scores of companies have tried to address these three markets but with limited commercial success. There are solutions for in-home video distribution using Wireless HD, WHDI, 802.11n and UWB –  and many of them commercially available. Several companies build outdoor WiFi gear for municipal deployments. Broadband can be provided in rural areas with a wide range of mature wireless technologies (EVDO, UMTS, LTE, WiMAX etc.) running in both licensed and unlicensed bands. Licensed spectrum is dirt cheap in rural areas. All this raises the question – is there enough demand for these applications and is  new technology needed to address them?

According to Ruckus, a company that builds WiFi gear with beam forming antennas, “white space will be ideal for creating “urban overlays” to higher-speed microcell Wi-Fi and macrocell LTE networks… perfect for offloading low bit rate “chatter” traffic, such as application notifications (email, presence lists, etc. generated by handheld wireless devices) from high speed cellular or Wi-Fi networks”. Though an interesting application, it is doubtful that carriers will add a new radio into their handsets to offload low bit rate chatter. The incremental cost of adding WiFi to a handset had to fall below $10 before carriers starting making it a standard feature in their smartphone lineups.

Brough Turner, founder of a a 802.11n based ISP called netBlazr and former CTO/founder of NMS communications, points out that white space spectrum being “beachfront” spectrum is based on 20th century technology, not physics. Brough, in other blog posts and presentations at industry forums, has argued that large amounts of spectrum at higher frequencies is significantly more valuable for offering broadband and connectivity than few 6 MHz channels in lower frequency bands.

Of course, those who have commercially viable ideas on what to do in this spectrum are not advertising them on the Internet.  I was recently reading the history of ISM bands on the website of Michael Marcus and at George Mason’s Internet Economy Project. It is notable that both WiFi and Bluetooth, poster child applications for ISM bands took off more than 15 years after this spectrum was opened up for unlicensed use. Spread spectrum, the technology that folks at FCC believed would be deployed in ISM bands was replaced by OFDM. None of the companies that were pioneers in the ISM band are in business today.  Plus, not all unlicensed spectrum creates billion dollar markets. Unlicensed PCS (UPCS), a 20 MHz band what was offered for unlicensed use in 1995, has no application to date. Still, in these relatively early days, it is better for all us to stay optimistic about the possibilities and keep our thinking hats on!

White Space Spectrum Becomes Available With Simplified Rules

White Spaces Officially Available for Unlicensed Use

On 9/23/2010, the FCC officially made unused TV channels (so called “white space” spectrum) available for unlicensed use and resolved several open technical and legal issues. Most importantly, in its Second Memorandum Opinion and Order, it eliminated the requirement that TV bands devices that incorporate geo-location and database access must also include sensing technology. This is a big change that simplifies the development of devices that use this spectrum.

The FCC had first voted to approve unlicensed use of white spaces in Nov 2008 and issued a “Notice of Proposed Rule Making”. This started a a consultation process that culminated in last week’s order. The Nov 2008 required unlicensed TV Band Devices (TVBDs) to not only incorporate geo-location and access a FCC-mandated database but also to include sensing technology to detect TV transmitters and wireless microphones. Fixed unlicensed white space devices (li.e. base stations) had to do spectrum sensing with a 10m high omni-directional antenna! These rules effectively made the spectrum useless for both – providing broadband access or local area networking (see posts from last year). Though I am skeptical of the FCC’s “Super Wi-Fi” claims, I do think the revised rules make the spectrum more interesting for both these applications.

TV Band Devices and a TV Band Database

The FCC order defines two kinds of TV band devices: Mode I (or personal/portable) and Mode II (fixed), and a TV bands database. Mode II (fixed) devices must either have geo-location capability (50 m accuracy) or be professionally installed. In both case, these devices report their location to a TV bands database and get a list of permissible channel on which these devices and the Mode I devices connected to them can operate. Mode I devices operate on channels provided by Mode II devices.

Simple White Space Network

Several companies, including Google, are interested in building a TV Bands Database. This database would contain information on channels used by protected services by geographical location. Details of this database are still being worked out and the FCC has to select database providers. The FCC does leave the door open for personal/portable devices that have spectrum sensing capability. Such devices can work without a Mode I device to form, say, a mesh network, and it will be interesting to see how companies that have done considerable amount of research in cognitive radios will react.

Not Much White Space Spectrum in Major Metro Areas

Still, the real barrier to the development of commercial white space devices is the availability of white space spectrum. FCC rules still bar the use of Mode II devices in both channels occupied by TV broadcasters and in adjacent channels. As a result, there are often just one or two 6 MHz channels available for Mode II devices in major metro areas. Mode I devices are allowed to operate in adjacent channels if the transmit power is less than 40 mw.

Spectrum Bridge, one of the companies that want to provide a TV band database, offers a free online service www.showmywhitespace.com that allows one to check the channels available for white space spectrum.

Super Wi-Fi. Really?

It is still not clear to me how white space spectrum will unleash super Wi-Fi. Compared to a Wi-Fi router, a TV band wireless router will have to share 6 MHz of spectrum with other similar routers in major metro areas and will require an in-building GPS receiver. Still, lots of other applications may be possible, and any tie-in with Wi-Fi helps in promoting new unlicensed band technologies (remember WiMAX?)!

An alternate way to use TV band spectrum

In the past, I have argued that FCC’s proposed rules for TV band white spaces make this spectrum unattractive for launching either wide area broadband services or local-area data/video services.  Limiting transmit power to 4W dilutes the coverage benefit of operating in 600 MHz (UHF) spectrum. Requiring CPEs to determine their location and connect to a central database makes CPEs expensive and difficult to install. The FCC is placing these requirements to ensure that white space devices can co-exist with terrestrial TV and all other existing services in these bands. Though these problems can be overcome with significant investment, the unlicensed nature of this spectrum weakens the business case to make such investment, especially since unlicensed spectrum with fewer strings attached is available. Bottomline – it will be years before TV band spectrum will be productively used.

All this raises the question – shouldn’t there be a better way to utilize this valuable radio spectrum? This afternoon, I was listening to a podcast on EconTalk with Thomas Hazlett – Professor at George Mason University and Chief Economist of the FCC from 1991-92, with stints at several other universities in between. Professor Hazlett argued (convincingly, in my opinion) that the FCC should pay all terrestrial TV stations to migrate to a free-to-air satellite system, give free satellite dishes to the 30 million or so Americans who still watch terrestrial TV, and auction the almost 300 MHz of spectrum that will be opened up.  Auctioning ~300 MHz of spectrum in the the 400-700 MHz band will more than cover the cost of the TV transition. If you decide to listen to the podcast, the discussion on spectrum starts after 44 minutes.

Just imagine the kind of wireless broadband services that could be deployed if ~300 MHz of spectrum in low frequency bands was made available. It would make wireless broadband a credible competitor to cable and fiber in the last mile, increasing consumer choice and lowering prices.  I am surprised that groups like the Wireless Innovation Alliance are not pushing for something as radical as this rather than unlicensed use of TV band white spaces.

White Space Spectrum Gone Missing? Just 1 Vacant Channel in LA, NY and SF

Opening TV White Spaces is supposed to bring us “WiFi on Steroids” . Really?

In one my previous posts, I had argued that the amount of white space spectrum available (as currently defined in FCC’s NPRM) is just not sufficient for wireless ISPs to launch services in urban areas.  But I was still optimistic about the use of white space spectrum for consumer electronics applications. So, when Spectrum Bridge launched a web-based tool to show the white space spectrum available at any location, I decided to use it to find the number of “white space'” channels available in downtowns of 6 major metro areas.

Here are the results:

• Boston (@62 Boylston St): 5 channels
• Chicago (@ N Michigan Ave): 9 channels
• Dallas (@901 Main Street): 2 channels
• Los Angeles (@138 S Central Ave): 1 channel
• New York (@ 1460 Broadway Ave):  1 channel
• San Francisco (@1298 Howard St): 1 channel

If Spectrum Bridge’s database is correct, it means that just one channel is available in downtown LA, NY and SF – three of the largest metro areas in the US.  This severly limits the use of white space spectrum for consumer electronics applications like home networking or in-home video distribution. Even a few towns/cities like LA, NY and SF are sufficient to increase return rate on a consumer electronics device by a few percentage points, which is terrible for the profitability of the product line.  Remember that white space devices also need geo-location capability as well, and they need to compete with mature 802.11 b/g/n devices. Overall, not great for the backers of CogNea standard (see recent press release from GeorgiaTech about goals of this alliance) or startups that want to enter this space.

Behind the White Space Database Announcement…

Yesterday, Google announced that it has teamed with Microsoft, Dell, HP, Motorola, Neustar and Commsearch to create a group to advise the FCC on how to create the “white space database”.  

The so called, “white space database” is a big sticky point in FCC’s Nov’08 Notice of Proposed Rule Making (NPRM) on white spaces. The FCC decided that virtually all white space devices need to have geolocation capability (i.e. GPS) and be able to access a database of protected radios services (such as TV and public safetey) before operating.  The only exception is personal/portable devices operating under the control of a fixed device.

 The FCC made this decision because in its tests of five white space device (WSD) prototypes last year, only Motorola’s prototype device correctly reported all occupied channels used by TV stations.  Since Motorola used a geolocation/database access feature in combination with spectrum sensing, the FCC decided that a database to protect existing radio services is essential.

Both requirements –  geo-location and ability to connect to a database over the Internet – are very problematic for commercial use of white space devices.  GPS does not work well indoors and consumers may not always want to connect to the Internet.  It means that if you bought a home router that worked in white space spectrum, you will not be able to set it up in your basement without an external GPS antena.  And that, you will not be able to use a “WiFi-on-steroids” network in ad-hoc mode. The list of applications that would be restricted goes on.

Further, this database of protected radio services does not exist today.  Creating it is not easy because it must include the “service countour” of every protected service.   This process has not started yet can drag on for a very long time if all the current users of this spectrum decide to drag their feet.

According to the FCC’s NPRM, “the database will be established and administered by a third party, or parties, to be selected through public notice…” Google and its partners in the white space database group are trying to influence this process to make sure that this database uses an open and non-proprietary format.  Of the two smaller companies in this group, Comsearch (a spectrum management company) is probably wants to be involved in establishing service contours, and Neustar (a clearinghouse) would likely want to adminster the database.

Even if the database gets established, the geolocation requirement will restrict the adoption of white space systems. The alternative is for the cognitive radio technology companies to prove beyond doubt that spectrum-sensing along is sufficient to protect existing services.  This process may take a few years (research, building prototypes, FCC testing, public review of results, final ruling), but is essential for widespread use of white spaces.

 

More on White Spaces: IEEE 802.22 and CogNeA

IEEE 802.22

People researching technology options for TV white spaces should read an article on the status of the IEEE 802.22 Working Group in the January 2009 issue of IEEE Communications magazine.  Launched in Nov 2004, IEEE 802.22  was charged with developing a standard for cognitive radio based air interface in TV white space spectrum.

 Before this article was published in the Communications magazine, information on the status of this standard was difficult to find unless you were a member of the working group.  Though I cannot do justice to summarizing this article in a blog post, here are a few key takeaways:

• The objective is to create a standard for fixed point-to-multipoint communication. No mobility (or portability) is envisioned
• The network architecture, MAC and PHY are derived from 802.16 (WiMAX)
• In order to work within TV white spaces, both the base station and CPE must:

           –  Support spectrum sensing with a 10m high outdoor antenna
            – Automatically determine its location using GPS (or Galileo)
           – Connect to a database of incumbent TV channels

• The standard is being designed to operate with one 6, 7 or 8 MHz channel in TDD mode. MIMO is currently not supported
•  Multiple 802.22 systems can co-exist in the same region as long as they transmit a “co-existence” beacon to identify each other.

 From a business point of view, this standard will be uncompetitive in any place other than the most remote rural areas, for the following reasons:

1. As proposed, this standard will offer much lower data rates than competing OFDM standards (WiMAX and LTE) because it operates in a 6 MHz TDD channel. Compare this to the 20 MHz spectrum that Verizon (and AT&T) have in the 700 MHz band for LTE deployment.
2. Unlike LTE or WiMAX, 802.22 does not support any mobile or portable operation. On the contrary, it requires every CPE to have a 10m external antenna for spectrum sensing.
3. The only advantage that 802.22 has is operation in a lower frequency band (500 – 700 MHz). This advantage is nullified by FCC’s requirement that even fixed white space devices have a maximum EIRP of 4 W.

The standard also seems to have some technical discrepancies, like it expects the CPE, which does not have an Internet connection, to connect to an incumbent TV database.  And its completion data is uncertain.  In any case, I would be surprised if there are any companies building equipment or CPE for 802.22.  If there are any, I would be really interested in hearing from them.

CogNeA

At this stage, it look like  most of the interesting activity in standardizing systems that utilize TV white spaces has shifted out of IEEE 802.22.  In December 2008, Philips, Samsung, HP, TI, ETRI and GeorgiaTech announced the formation of a new group called the Cognitive Networking Alliance (CogNeA). CogNeA aims to drive the definition and adoption of an industry-wide standard for low power personal and portable devices to operate over TV White Spaces in the Ultra High Frequency (UHF) television bands. Notice the focus on personal/portable devices rather than providing broadband.

This group does not have a website yet, though they shared some of their plans at the Software Defined Radio Forum’s meeting in San Diego last week.  If you are following this space, add CogNeA to your Google Alerts!

White space spectrum unlikely to enable new ISPs

FCC’s 11/14/08 announcement to open up TV-band white spaces has been hailed as a great step forward for wireless inovation, and countries around the world seem eager to follow in FCC’s footsteps . There is lot of excitement about this spectrum because it has excellent propogation characterstics.  In my opinion, even though this spectrum may enable some new consumer electronics applications,  this spectrum is unlikely to create any new service providers.

Some background… White spaces exist because TV broadcasters have to do frequency planning. A channel that is used in Los Angeles cannot be used in Orange County or San Bernandino and vice versa. After the DTV transition is over, LA will be using 11 channels (7, 9, 1, 13, 28, 31, 34, 36, 41, 42, 43) in the VHF and UHF band, leaving 40 channels unutilized.

After years of investigation, FCC decided to allow unlicensed devices to operate in this spectrum subject to certain conditions.  Here is a quick summary of FCC’s Notice of Proposed Rule Making (NPRM) on this topic. 

• Unlicensed TVBDs are expected to protect all services that operate in TV bands today. This includes broadcast TV, wireless microphones, medical telemetry, radio astronomy and land-mobile radio systems
• Unlicensed TVBDs are not allowed to operate in any channel being used for Broadcast TV (i.e. generate no co-channel interference)
• Based on rules for adjacent channel operation, maximum transmit power and spectrum sensing, unlicensed TVBDs can be of two kinds: Fixed and Personal/Portable.
• Fixed devices
  • Use channels 2-51, with the exception of channels 3, 4 and 37
  • Maximum transmit power (EIRP) of 4 W with directional antennas
  • No adjacent channel operation with TV
  • Support spectrum sensing, and detect signals as low as -114 dBm
  • Determines its location; accesses a database of existing services
• Personal/Portable devices
  • Use channels 21-51, with the exception of channel 27
  • Maximum transmit power of 100 mW. No directional antennas allowed
  • Maximum transmit power limited to 40 mW in channels adjacent to TV
  • Spectrum sensing, location determination, and access to database of existing services required, unless device is operating under control of a fixed device

Notice that the FCC does not allow fixed devices to operate in adjacent channels and requires portable devices to reduce their peak power. This restriction reduces the amount of “white space” spectrum available to a new ISP considerably. Going back to the LA example, once we removes channels 2, 3, 4, 37 and all adjacent channels from the list of available channels, a “white space” ISP has only16 channels, not 40. Though sixtenn channels provide 96 MHz, an ISP must be prepared to share it with other users like wireless microphones and personal/portable TVBDs.

More white space spectrum is available outside major metro areas but these areas also have abundant and inexpensive PCS, Cellular and AWS spectrum. Despite its better propogation characterstics, Unlicensed TV band spectrum does not cover much more than licensed PCS or AWS spectrum because the transmit power of unlicensed TVBDs is limited. As a result, an operator/ISP that deploys a proven 3G/4G technology in licensed bands is guaranteed to do better than on ISP that uses yet-to-be developed “white space” technology.

Any ISP using it in a large metropolitan area will also have to contend with three to four established competitors (cable, DSL, WiMAX and LTE). Technology required to build a commercial white space system is barely on the drawing board. The IEEE standard (802.22) that is being developed for white spaces is far from completion and in its current state does not meet FCC rules.  Even if lots of vendors get behind this standard, it will be at least four years before commercial base stations and devices can be available at reasonable price points. By that time, it is debatable how much unmet demand will exist, in the US or outside.

I think things can be more interesting for personal/portable devices, but that is the topic of another blog entry.