John, I think we've already seen evidence the cable modems aren't a good solution for the internet. Even the latest TCI initiative talks about services like local weather, sports and home shopping, not full internet access. I think it's also significant that the new WebTV Plus box from MSFT/WebTV uses a 56K POTS modem for Internet access and upstream data, and a 1 Mbs modem for cable. (I believe they use the vestigial sideband of the carrier for cable content). The cable modem will be used primarily to access local content, ie; download a movie to the hard disk resident on the WebTV Plus system, or view a TV program and TV content related data simultaneously on the TV. Even this isn't very impressive. Considering it takes hours to download a movie, how do you suppose a lot of people are going to download different movies at different times?
The problem for cable is that even with fast two way modems, how does one get access to content that doesn't reside on the local server? If I want to access a site in Taiwan, how does the cable provider get to it? Connect to the web on the switched phone network? That's my guess. My suspicion is that the proposed new cable set-top boxes will eventually provide both cable modem and POTS modem if they're serious about providing full internet access. If so, how will they price this service? For those that want it they'll probably need to sign up for both the enhanced cable services and an internet provider.
From what I've seen of the TCI initiative they are talking about local services, enhancements to TV programming, and guess what, 500 channel cable. Seem's to me we've been hearing about 500 channel cable for almost ten years now.
The other problem is that cable is more analogous to a bunch of LAN's than to a web. Put a metropolitan area on cable and I think you'll see real contention problems. While cable access may be fast when there are only a few thousand subscribers, what happens when there are hundred's of thousands? Wouldn't surprise me if POTS is faster than cable a lot of times.
Note the target price for these boxes is $300. Probably need a cost of less than $200. How much profit is there for the hardware and software participants? Since TCI wants this to be an open system, can Intel really compete with Cyrix/NSM, AMD and IDT on the CPU content? Yes, Intel is proposing that a PII provide multiple functions, but this is real low margin commodity business.
How many local providers will actually implement this system? Some major changes to the head-end. Will these be implemented before the consumer buys the set-top boxes, or will the provider wait until there is sufficient demand for the service?
For true high-speed internet access my bet is on XDSL. Think it will happen long before cable.
What TCI proposals I've seen don't come close, ("The poor man's internet.")
Just my opinion. |
