DOCSIS 3.0 Q&A: Peter Percosan

Peter Percosan Title: Executive Director, Broadband Strategy, Texas Instruments Digital Connected Home

Background: Persosan has been strategy director since TI’s acquisition of Libit Signal Processing in 1999. At Libit, he had served as VP strategy and business development. Previously he was VP strategy marketing at Lysis SA and VP engineering at Ariel Corp. Earlier this year we spoke with him about PacketCable 2.0 Your news this week concerns Texas Instruments’ introduction of DOCSIS 3.0 cable modem chipsets that are supporting CableLabs certification wave 56. What’s the bottom line message?

We set a stake in the ground, executed, and we hit it, and this is real.

Has Cert Wave 56 started?

I’ll have to leave it to CableLabs to discuss when it actually starts, but it’s fairly public that everybody’s gear has to be in Louisville (Colorado) this week. Then the party begins.

How long will it take?

This is going to take a long time. A normal wave is in the order of eight to 10 weeks, and that is nearly 100-percent automated. We’re definitely not at that point. Plus, the spec is way more complex; there’s a lot more testing to it. I think the goal is to have results by the end of the year. But Cert Wave 56 is not just for 3.0. The 56 results for 2.0 modems should come the first week of December. I’d have a heart attack if we were done (with 3.0) by that point … but it would be a good heart attack.

What’s involved with automation?

There’s a science to this. That’s basically CableLabs’ value-add, and one of the goals of the contributing engineers that companies like TI locate there is to come up with ways to automate the whole testing process. The first step is generating tests; the second step is automating. There’s a huge amount of tests that are going to be automated for 3.0 because a lot of the testing for 2.0 is already automated. As for all the new features that we have been developing for months now, the emphasis has been on developing the tests, not on automation. So automation will come over time.

What else can you say about TI’s work over the past several months?

The energy that TI has spent breaks down into two major vectors. The first vector is development of the silicon and developing the software that runs on the silicon, and doing that under the umbrella of working with our major customers who are going into the wave.

The second effort has been on tests and interops. CableLabs has had a number of interop sessions where various CMTS and CPE vendors come in to participate and run against these tests that we have been writing, to make sure that everybody is interpreting the spec the same way and to make sure that simple things don’t go amok.

Remember that this is a first wave. We’ve been through first waves before, like the first 2.0, the first 1.1. One thing I really want to underline on this "first" is our level of feature completeness and stability is higher than we’ve ever had on any previous DOCSIS node.

Does that translate into long hours?

It translates into a lot of long hours for a lot of people. Many 16-hour days every day, not just weekdays, but through the high holidays. (The design engineers in Israel) have shown a level of commitment that really makes me proud. They really want this to happen. It’s not just a business for them – it is intensely personal. They really want to win.

And winning means passing, but also starting a new industry?

It’s a 3.0 industry. It’s not just about 160 Mbps downstream; it’s not just a speed upgrade. This team has taken a taken a personal ownership in the whole transition that the cable industry is going to go through.

Anything comments on any other chip vendors?

I have to assume that they’re taking it pretty serious now because we’ve been pretty vocal about how serious we’re taking it. I don’t take them lightly, that’s for sure.

How does the international market look?

Europe is going to be a big market for us. UPC is taking this very seriously. It’s been very refreshing. We’re trying not to focus on just U.S.-based companies, although Comcast has been an incredible partner.

What does it mean to be a good operator partner?

At broad strokes, it’s been more than just encouragement or someone in purchasing saying ‘I hope that 3.0 shows up some day.’ It’s not just Brian Roberts saying ‘3.0 is important.’ It’s been rolling through to various levels of the company to real execution. That’s been refreshing.

Many people don’t quite appreciate that vendors are the R&D arm of the industry, do they?

That’s an absolutely accurate assessment.

So now it’s time to see what happens with 3.0, isn’t it?

Let’s see what happens. The party is starting. Now it’s time to see if all the love translates into real market momentum and business.

What about other options for cable operators, such as PON architectures or 1 to 3 GHz spectrum upgrades?

1 GHz, for sure. That almost falls out for free, if they’re doing a rebuild. 3 GHz, not so sure. One thing about GPON is that it’s very much like the small node architecture for HFC. Are operators looking to pulling fiber in a point-to-point architecture? Yes, it’s their job to look at it. But I think of DOCSIS 3.0 as a 10-year node. It’s not just about 160 Mbps down. It’s business rules applied to a GHz plant and, combined with switched digital, gives operators a giant amount of runway.