Stream Dream

Forget voice over Internet Protocol (VoIP): for
UTStarcom, it's TV over the Internet. The 13-year-old company has built a
business based on IP networking, service and support that is expected to have
revenue of $2.8 billion in 2004. The company recently rolled out its mVision
system, an IP-based video-on-demand system that the company says gives access
to 100,000 hours of content without actually storing all of the content in a
viewer's set-top box. So far, no customers have signed up for the service. Jeff
Paine, vice president of strategic marketing for UTStarcom, spoke with
B&C's Ken Kerschbaumer about a market that faces its
own set of challenges and opportunities.

How do you view the TV-over-IP

In the short term, TV-over-IP's market really comes from the traditional
telcos. Those are the guys who can deliver services over either copper DSL or
some flavor of fiber to the home. And what's going to make it compelling for
them is, they have serious subscriber problems. They're losing land lines.

About 10 months ago, the telcos had an epiphany where they realized that
IP services were where they needed to go. And they realized they could go after
additional IP services, particularly in the entertainment/media content space
and get some nice numbers out of it [using their DSL service]. The corollary
was that, if VoIP could deliver about $10 a month, then entertainment content
delivered over copper or fiber becomes $30 to $40 a month.

You say your system can store 100,000
hours of content in a single rack at the headend.

We're not actually storing the movies, per se, at the edge. We're an IP
company, and we're using switched video delivery to the edge. So we store the
first four minutes of every piece of content on the edge server. And the minute
the user orders the content they want, it downloads that four minutes. And we
have a smart protocol that starts shipping more segments out to the system.

Also, our system allows the content to be stored at the movie studio. We
have a media-location registry that knows where everything is and where people
are. So we have a way to provide hundreds of thousands of pieces of content at
the same time.

What about the other cable

Depending on the country and rebroadcast regulations, we could not only
stream content over the network but store content in the network. So it's
feasible to have premium- and standard-cable programming stored as well as
weeks' worth of local broadcasts.

So when do you expect to see some

We have a substantial deployment under way right now, but we can't talk
about it. I think this stuff really starts to roll out in the middle of next
year. It's going to happen everywhere, but it's just going to take a while for
the U.S. telcos to figure out the economics.