Archive for October, 2009

Hello, Open Network Architecture… We’ve Missed You

Wednesday, October 14th, 2009

Bam!  The Berkman Center report on broadband is up on broadband.gov.  And it’s a screamer, as these things go.  I’m still working my way through it but it looks like the big news is a call to return to unbundling and open access in U.S. communications policy, supported by actual evidence.

The report charmingly notes at one point:  “Talking about ‘unbundling‘ or more broadly open access in the United States today is a bit like wearing bellbottoms or talking about a national healthcare system” (p. 77)

Ten years ago I was a lowly research assistant on a project that made the case for open access rules in terms of innovation: “Access and Innovation Policy for the Third-Generation Internet” (link to pre-print).  Jean Camp once referred to this article as the most useful analysis of Internet open access ever written (thanks to Bar, Cohen, Cowhey, DeLong, Kleeman, and Zysman for writing it).  It led to testimony before the FCC but no open access rules.

Unbundling basically means to force things that were sold together to be priced separately.  E.g., force cable companies (like the former TCI) to allow wholesale use of their monopoly cable plant so that other companies can be Internet service providers without building their own cable network.

Anticipating all that network neutrality stuff that people are so exercised about these days, my job was to find information about what the cable companies were up to in the provision of Internet access, and also to try to estimate how easy/hard it was to switch from one form of Internet access to another.  In the annual report for then-cable-company TCI (public because it was submitted to the SEC) we found the company bragging about providing differential service to content when it got what was essentially a kickback.  If I recall, some games were faster on TCI’s Internet service because TCI got a percentage — the differential speed had nothing to do with the quality of the game or software or network.

The idea behind this research was: if there is any of this tricky stuff going on and it costs a lot to switch types of Internet service (from cable to DSL to fiber etc.) then we have a solid argument for unbundling and open access to promote a diversity of ISPs.  With a lot of ISPs hopefully some of them won’t behave so badly.

Ever since the computer inquiries a lot of mainstream researchers in telecommunications with a U.S. perspective have managed to portray unbundling as a crazy and unworkable idea that they complain is focused on unworkable distinctions and definitions.  I recall a witticism from ten years ago suggesting that an unbundled Microsoft would be forced to sell other software makers individual access to the cursor, and the blue window that surrounds each application, and the text in that window, and access to the hard disk, and so on…  but this misses the point.  As the Berkman report makes clear we have evidence that unbundling has been defined and it seems to be working in other countries.

Bellbottoms are back in fashion by the way.

Bad Behavior has blocked 1364 access attempts in the last 7 days.