Comcast or Comca$t?

July 5, 2008 10:51 AM

What are the costs of doing business for a large ISP?

All this talk about bandwidth throttling, downstream capping, and miscellaneous connection meddling among the major ISPs has made me wonder why the market isn't producing an alternative to the major internet providers. In the old days of dial-up, there were often dozens of small-scale ISPs per city, while today a few major companies - Verizon, Comcast, etc. - dominate the national market.

Here are my questions:
* What are the operating costs of a business that provides high-speed internet connections? What's the relationship between those costs and the fees charged to customers? Is the argument (made by ISPs) that rising costs will mean increased fees credible?
* What factors are preventing new competitors from entering the market? What's to stop an upstart ISP from renting Google's "dark" fiber optics?
*And finally, what are some potential solutions that a new company might employ to get around today's existing constraints?

posted by awenner to computers & internet (7 comments total)

Verizon, Comcast - whoever you pick, they're the ones running the pipe into your house, be it fios or coax. Back when there were a gazillion ISPs, they were all using the phone lines for pipes. Note that most of the remaining "alternative" ISPs - around me, I know about speakeasy and cavalier, just offhand - are DSL, which allows them to, yep, piggyback on the existing phone-line infrastructure.
posted by Tomorrowful at 10:55 AM on July 5

1. Operating costs - you need to lease enough space in a telecom hotel to house your routers, billing and authentication servers. You also need to lease bandwidth from one or more major backbones. Finally, you need some way to deliver bandwidth to your customers. Oh, you also need to account for marketing and operating costs.

2. Anyone can enter as a CLEC, leasing last-mile copper from the ILECs. However, the ILECs don't have much motivation to improve their equipment, so don't expect anything particularly fast.

2b. If google is willing to light up their fiber and lease some of it to you, then go ahead and write a contract with them. It will be much, much, much easier to purchase bandwidth from an existing backbone provider. I wouldn't even bother considering the google option.

3. The easy way to start up a small-scale ISP is to lease last-mile connectivity from your local ILEC. Another way to go would be with your favorite wireless technology.


The backbone business is highly competitive. Bits are very cheap. Getting from the backbone to the customer is tricky. If you last mile connectivity from an ILEC, you are offering the same connectivity everyone else is. You have to market the hell out of your offering or drop the price. Alternately, developing a new last mile solution will be risky and capital intensive.
posted by b1tr0t at 11:17 AM on July 5

To be clear, I'm not personally looking for a business model. I'm am interested in understanding how, if, and when more attractive alternatives to the existing large-scale ISPs will appear.
posted by awenner at 11:50 AM on July 5

What do you consider unattractive about the current large-scale ISPs?

They serve most people's needs, at an attractive price. There are numerous small specialty ISPs for people with more unique needs. What market demand do you see that is not being met?

Sure, you can get a higher bandwidth link in Tokyo, but what is the typical consumer going to do with it?
posted by b1tr0t at 12:43 PM on July 5

One place to look for where attractive alternatives to the existing large-scale ISPs are to other large tech companies who aren't ISPs.

Both Microsoft and Google realized, perhaps before the ISPs themselves did, that bandwidth is a complimentary good to their businesses, and that as such, their core businesses benefit from bandwidth being inexpensive. They also realized that the overall lack of competition among regional broadband ISPs put the ISPs in a position for "rent-seeking," by trying to extract cash for access to the ISPs customers.

As a result, they've been looking at ways to keep the last mile competitive. Both seem to be taking a two pronged approach combining both public policy (ie lobbying) efforts, with technology development. This can be seen quite visibly in Microsoft & Google (and others) teaming up on the development of "whitespace" networking. Whitespace networking makes use of relatively large swaths of bandwidth that serves as padding between TV channels. This padding is important for high powered transmission using technology developed in the 40s and 50s, but modern data radios working at lower power can use that bandwidth for high speed transmission without causing TV interference. Or that is the theory. Google and Microsoft have developed equipment to demonstrate feasibility and are pushing the FCC to approve it.

It can also be seen in Google's strategy in recent spectrum auctions where they were willing to ante up billions in order to make sure that whoever bought the spectrum would open their networks to "any approved device."

I think Google's municipal WiFi project in the bay area is also in this vein.

I don't think either company has ambitions to actually go into business as an ISP, but they are demonstrating that they could should existing ISPs get too greedy. In doing so, they also strengthen the hands of state regulators, and municipal governments. Some cities seem to regularly threaten to create their own fiber optic networks in order to spur the incumbent phone and cable operators to upgrade infrastructure & keep prices reasonable. Any technology and policy that makes it easier for an upstart helps keep the existing ISPs "honest".
posted by Good Brain at 12:55 PM on July 5 [1 favorite]

What others have said regarding the "last mile" is absolutely correct - this is where all the costs (as well as most of the bottlenecks) occur.

The reasons you do not see the market producing a viable broadband competitor are twofold:

1. (as others have mentioned) the cost of leasing last-mile connectivity is prohibitively expensive in all but the largest markets,

and

2. Only the telephone companies (or, increasingly, THE telephone company, aka Verizon aka formerly known as the phone company the government forcefully dismantled, re-formed) have common-carrier status. What this means is that the phone company has to allow access to their infrastructure by third-party business. Cable companies, as the only other available/dependable last-mile solution, are not common-carriers and thus do not have to allow anyone else access. This is why you cannot shop around for cable internet access the same way you can for DSL or dialup access.

As for when you will see the market produce a viable alternative? Whitespace networking, if it gets approved, is a good candidate, but a better candidate will probably be the C-block of the 700 Mhz wireless spectrum, or (less likely in my opinion) WiMax. Only time will tell.
posted by namewithoutwords at 3:05 PM on July 5

Although I haven't look in a while, Public Knowledge might have some good information on the economics of broadband. Check their papers and blog.
posted by PueExMachina at 10:15 PM on July 5