Communications

Access to right-of-way and it's impact on Broadband cost & performance

For my presentation at eComm 2008 in March, I searched out the cost and availability of 100 Mbps symmetric Internet Access in various locations.  At that time (1Q08), the ranking and approximate monthly cost (in US $) was:

Stockhom        $16

Tokyo              $25

Seoul               $32

Hong Kong       $35

In each case, 100/100 Mbps symmetric service was widely available in the city and in many other parts of the country, (but I only obtained costs for those specific cities).

Now I have a new data point to add:

Bucharest        $11
Image via WikipediaFor 9 Euros per month you actually get a quad play service:  CATV + FTTB (FastEthernet 100 Mbps) + fixed voice + mobile with some hundreds of minutes included.  The only catch is connectivity to the rest of the Internet is subject to congestion.  Here's a mid-afternoon measurement between Bucharest and Frankfurt which shows 43 Mbps upstream, but less than half that downstream.   Thanks to Catalin for the detail and the measurement.

Access to the right-of-way

It's interesting that the best deal is in Bucharest where there is open access to public rights-of-way, as I blogged last month.

The second best deal is in Stockholm where dark fiber is widely available to all comers.  Stockholm's abundant dark fiber is mostly due to Stokab AB, a government sponsored corporation originally chartered in the mid-90s to install dark fiber and make it available at break even costs.

TV White Spaces - Fundamentals misunderstood

The FCC's recent decision to allow secondary use of the TV bands by license-exempt devices is a political breakthrough we can all rejoice in, but the widely claimed technology benefits are at least partly bogus.

Here's a typical claim:

This is the same spectrum we have referred to as "prime beachfront property."  < ... >  These lower frequency signals lose less power as they travel, hence requiring fewer base stations to cover an area. They also feature far better penetration characteristics, which is key to improving indoor operation.

This is true of deployed systems today, but for engineering reasons, not due to the physics of electromagnetic waves.

Today, wireless technology is advancing at such a pace that higher frequencies, for example at 5 GHz, will be more useful than TV White Space, perhaps within the time it takes TVWS equipment to get widely deployed -- certainly within 5-10 years.  What's happening is MIMO and advanced MIMO driven by emerging WiFi (802.11n), WiMAX and LTE systems.  As these systems get developed and deployed, signals at 5 GHz will achieve similar or greater range than TVWS, similar or better building penetration, lower costs and substantially higher carrying capacity.

The reason buildings appear to attenuate high frequency signals (like 5 GHz) more than those in the TVWS bands is, as radio signals hit the different materials that make up walls, windows, furniture, and other building elements, they scatter (due to diffraction and reflection).  With conventional receivers, only the primary signal is detected while scattered signals appear as noise, degrading the received signal.  Shorter wavelength signals are scattered by smaller objects and thus by more of the building's elements.  The higher the frequency the shorter the wavelength, so it's higher frequencies that experience more scattering.  But if you have multiple antennas and multiple receiver channels (MIMO), you overcome scattering.

Indeed with MIMO, the tables are turned.  MIMO systems use multiple antennas, typically separated by several wavelengths, or beamforming with many antenna elements each separated by 1/2 wavelength, but that's a problem for TVWS sysems.  At 5 GHz, a wavelength is 6 cm or just over 2 inches, but TVWS wavelengths range from 1.4 meters (55 inches) to 5.6 meters (over 18 feet).  Separating multiple antennas by several wavelengths is a problem at TVWS frequencies.  It might be feasible at a fixed base station but it's too bulky for a residential device and completely nuts for a laptop computer or a mobile phone.  So expect MIMO to be widely deployed at 5 GHz, but not so widely or not at all in the TVWS bands.

There are many other reasons why physics favors 5 GHz over the TVWS bands, but MIMO is the one place where engineering is catching up with physics as we speak.

Secondary Use of spectrum ― TV White Spaces just the beginning

Yesterday, the US Federal Communications Commission voted to allow qualified devices to operate on a license-exempt basis in unused portions of TV channels 2-51, spectrum commonly referred to as "TV white spaces."  A lot has already been written about this move which was supported by Intel, Google and others in the White Spaces Coalition, but vilified by TV broadcasters.

What missing in these discussions is the bigger picture.  Almost all radio spectrum is under-utilized and yet, except for a few slivers, it's all been assigned to someone. Meanwhile today's radio technology is vastly more efficient than most of what has been licensed.  What's more, we now have the potential to share wireless spectrum in ways that potentially render the whole idea of fixed frequency assignments obsolete.  But those who have licenses under the existing regime represent strong vested interests.  The only politically acceptable approach is to open up "secondary use" while protecting legacy licensees.

The first step was getting access to the 3650-3700 MHz spectrum, something that happened last year.  Yesterday's move was a much bigger step, because the legacy licensees in the TV bands are so vocal and so well connected in Congress.

Now we need similar access to all the rest of the spectrum!

To elaborate...

Am I too optimistic about TV White Spaces?

Some have complained the white space rules are too restrictive and will never allow the technology to reach it's promise.  But these rules are subject to change as technology improves, as initial fears are overcome and as the growing "secondary use" community gains critical mass and political clout.  Such FCC rule changes are routine and are normally completed without the mass of publicity that surrounded yesterday's decision.

Spectrum under-utilized

Today, almost all radio spectrum has been licensed to someone although most spectrum is unused most of the time, even in major metropolitan areas.  Obviously there are a lot of unused TV channels in any one city, as illustrated by these energy measurements in Washington DC's Dupont Circle:

but most spectrum, particularly from 1 GHz to 10 GHz is under-utilized:

Radio technology has been improving at least as rapidly other electronic technology and yet most spectral assignments were made decades ago and mandate the use of specific technologies that are completely obsolete.  What's worse, it now appears that the whole concept of assigning specific frequencies to specific users was just a result of limitations of early- and mid-20th century technology.  Today, and more so in the future, there are orders of magnitude of additional performance that can be obtained from wireless technology.  But in almost all cases, new technologies are blocked by existing assignments and their associated vested interests (like the National Association of Broadcasters in the case of TV white spaces).

The Visual Analogy

To get a sense of the potential, consider the visual spectrum (light waves are just very, very high frequency radio waves).  I don't need a license to turn on a light bulb.  No one has licensed the sun.  And no one has licensed the lightwave receivers that are our eyes together with our visual cortices.  And yet we receive an enormous amount of information visually and we do so despite others in our immediate vicinity operating with exactly the same frequencies.  Today's radio receivers don't approach the efficiency of the human vision system but we now understand that they could, and presumably will, in due course.

Obtaining secondary access to more spectrum

The technology that has been developed for the 3650 MHz band (see IEEE 802.11y for example) and the technology that is developed for the TV White Spaces, can be reused in other bands.  Already there are advocates for license-exempt access (as a secondary use) to the 800 MHz stretch of spectrum from 3400 MHz to 4200 MHz.  And, many other bands are candidates for license-exempt secondary use.

Let's take pleasure in the success in the TV white spaces, but resolve to go after all the white space across all the spectrum.