DSL Prime: Technology Advances

From VDSL to IPTV, DSL providers have huge technology choices to make, and DSL Prime says that, as usual, some public companies are concerned only about the short term and the bottom line.

by Dave Burstein of DSL Prime and Future of TV [February 18, 2005]

VDSL2: 40Mbps loop lengths of 2,500 ft, 25Mbps at 4,000 ft
SBC releases performance data
Glen Campbell of Merrill, one of the best informed analysts, just wrote an important report on SBC IPTV after a visit to the labs. I've been begging chip vendors for VDSL2 data, because two world class designers have been skeptical of the 25 meg listed at 5,000 feet in SBC's and Bell Canada's original hopes. The chip tested is almost certainly Ikanos, the only chip shipping announced as "VDSL2," although last issue DSL Prime reported sightings of samples from Conexant. Real world performance from 2,500 to 6,000 feet is crucial for North American video plans, so I'll be looking for every result I can get.

Respecting SBC
Campbell concludes "Management has thought through and progressed many of the trickier implementation issues." SBC intends, per Campbell:

• a single TV set top box for multiple TVs
• installation times to decline over time from around four hours to two hours
• to soon decide key vendors
• whether to use MPEG 4 AVC or Microsoft Media Player VC1
• allocate 9.3 Mbps to a single HD channel, more realistic than their previously announced 6 Mbps. Standard definition channels will use 2.1 Mbps.
• allocate only 1 Mbps to DSL downstream, with some provisions for peaking beyond that
• VoIP and Wi-Fi built in to every gateway
• vacant copper facilities will be used to line-power DSLAMs.

One issue they've apparently resolved is the impracticality of the originally announced 6 Mbps down DSL data rates. Campbell's reporting pre-empts much of an item I've been working on, Respecting SBC, noting that much of their plan has possibilities. I need to clarify what I've been writing about SBC, because other reporters drew from my reporting that SBC's FTTN would fail in a business sense.

I believe we won't see how the business issues will play out at least until 2006 or 2007.

SBC is set to spend about $10 billion less than Verizon plans for fiber, an attractive saving. If that saves enough to allow SBC to price significantly cheaper, that could work. The key will be making real the $99 monthly price for the triple play Ed Whitacre discussed with Jon Van. SBC is hoping strong marketing, execution, and price discounting will win market share from cable and prevent losses of voice revenues. Will low price overcome data rates far slower than cable, or a customer's expectation to TiVO one HD program while watching another, impossible with a single stream? Will all the technologies come together to serve millions in good time? We'll see.

DSL Prime isn't predicting a failure of SBC's business, and certainly hasn't written anything to justify a sell sign on SBC stock, still $20 billion cheaper than Verizon's. Below, reporting on whether SBC will respect the open Internet. The only way I can imagine SBC giving 9.3 quality Mbps to alternate video providers will be to accept and forward individual outside streams at the headend. It would be a good thing.

Copper Mountain is Tut's new video box
Tut picks up pioneer for chassis, cash balance
"We love our CM DSLAMs," a CTO told me two years ago. "They never need repairs." Copper Mountain did everything right in 1999, selling thousands of DSLAMs to CLECs, with ahead of their time IP features. In the middle of the boom, when companies went public with losses, Copper Mountain established profitability early and had a market cap in the billions.

When the CLECs were killed, however, sales plummeted. Rich Gilbert took advantage of the advanced IP features of their later designs to claim major mindshare as an advanced BRAS and network controller, a product the Bells were hotly interested in. But the hundreds of million dollar Bell contracts were never issued, as Lightspeed and FIOS supplanted the previous network planning.

The speed of the platform, fully tested and NEBS compliant, make it ideal to deliver the forward error correction video system CEO Sal D'Auria announced at SMPTE two weeks ago. That's a solution desperately needed by many telcos deployments plagued with the dropouts of a typical IP routed network. Much of SBC's Lightspeed spending is on a new, high-reliability video backbone. Forward error correction is a plausible solution where a network upgrade doesn't make sense; every major carrier is looking at it.

My original assumption, when Tut announced they would retain CM engineers and advance the system, was that they would offer BRAS and DSLAM systems, which would put them in conflict with key partners like Paradyne. That was mistaken; the CM gear can be adapted to directly support video instead.

Tut is getting an excellent deal, because CM's cash will cover much of the $10 million purchase price, which is paid in stock. That will reduce leverage, critical for Tut's hope of winning larger contracts.

Copyright 2005 Dave Burstein.
The DSL Prime Newsletter is reprinted with permission.

"The power of the printing press belongs solely to those who own the presses"
—A.J. Leibling

The Internet is the cheapest printing press ever invented.