Switching Things Up on Channel Delivery


Some in the cable industry are moving towards making the switch to switched broadcast video (SBV) services, aided by new data from a pair of tests conducted last summer.

The results of those experiments were presented two weeks ago at the Society of Cable Telecommunications Engineers’ Emerging Technologies conference in Huntington Beach, Calif.

BigBand Networks Inc. participated in both tests, designed to develop mathematical formulas for how much a cable system could “oversubscribe” quadrature amplitude modulation units to achieve the bandwidth gains SBV provides.

In a SBV architecture, a certain number of channels are switched at the service-group level (usually 2,000 homes), which means that channels only travel the final distance to the home when the set-top tunes them in.

BigBand had launched initial SBV tests in 2002 and 2003 to work out any technical bugs.


The tests last summer — one in Cox Communications Inc.’s Tyler, Texas, system, and one in an undisclosed system — aimed to determine how many channels could be placed in SBV format.

The quick answer: It depends.

In Tyler, Cox provided 40-stream capacity for 60 channels to 603 digital subscribers. But in any given week, subscribers watched 50 of those 60 channels, providing little apparent gain from using SBV.

In the other test, 100 streams were set aside for 171 channels for 915 digital subscribers. Only 67 channels were ever used, representing a bandwidth savings of 67%.

The tests illustrate the difficulty in predicting how many — or which — channels work best in a SBV environment. It’s clear that any system considering SBV would need to collect large amounts of viewing data, cross-referenced across each node, to determine the best mix of channels. A channel that’s lightly viewed in one market might be popular elsewhere.

In Tyler, Cox used a Pioneer Corp. guide on a Motorola Inc. set-top in a 550-Megahertz plant.

It appears that a large portion of Cox’s more-popular digital channels were placed in the switched format. The 60 channels — including 32 premium feeds, six channels from Discovery plus other digital networks like TV Land, ESPN2, Sci Fi Channel, GSN, Fox Sports World and ESPNews — were sent the final distance to the home only when a subscriber tuned to them.

BigBand conducted engineering tests of SBV with Outdoor Life Network, Biography Channel, Do It Yourself, Fine Living, Toon Disney, Lifetime Movie Chanel, Bloomberg Television and Fox Movie Channel.


In the Tyler test, the number of programs viewed remained fairly steady — between 35 and 50 on any given day — despite the fact that between 95 and 170 set-tops were tuned in at any one time.

As 50 of the 60 channels were watched at any one time during the week, “this does not represent a large bandwidth savings,” said Cox ITV systems engineer Nishith Sinha.

But there were extenuating circumstances. When a set-top box changes channels in an SBV system, it sends a “tune-in” message to acquire the new channel information and a “tune-out” message to indicate it’s no longer watching the previous channel.

In the Tyler test, the “tune-out” data was not captured.

“Hence, the total number of channels used is artificially higher than normally would be expected,” Sinha said.

Cox also included more popular channels and no pay-per-view selections. “Substantial savings in simultaneous programs viewed is obtained when less popular channels are included in the mix,” Sinha said.

In the larger test, the 750 MHz system used Scientific-Atlanta Inc. set-tops. It carried a number of traditional digital networks, like The Outdoor Channel, Disney Channel, A&E Network, ESPNews, GSN, Toon Disney, Do It Yourself, MTV2, Noggin and Discovery’s digital networks, plus a host of premium and sports PPV offerings like “NBA League Pass,” “NHL Center Ice,” “ESPN GamePlan” and “NASCAR In Car.”

A much greater percentage of channels in this test were PPV channels.

In this trial, the average number of channels viewed ranged from 45 to 60, topping out at 67 channels on one day.

The corresponding number of viewers ranged from 85 to 120. The 67 channels viewed across 171 total, results in a savings of 60%, aided by the capture of “tune out” data, compared to the Tyler trial.


If those results held over a longer period of time, this system could set aside 70 streams to handle viewing for these 171 channels — cutting in half, for instance, the number of QAMs currently in use.

Those QAM savings can be applied to HDTV or other new services.

BigBand chief technology officer Ran Oz said he believes several MSOs will make SBV deployments in 2005, as operators look to save bandwidth. The deployment of digital simulcasting makes channel space even more important, and could spur SBV deployments, he added.

“Switched broadcast-video architecture is very, very similar to VOD,” Oz said.

The two main costs for SBV are the added QAMs and switches, he said. But many operators are starting to deploy edge QAMs for VOD, he said.

And BigBand has built a switching element into edge QAMs, reducing latency issues, he said. “That’s what we do today,” Oz said.