As cable services converge, so converge the myriad projects at CableLabs. One of the key people tasked with ensuring that they all work together cohesively is Ralph Brown, the newly-named chief technology officer of CableLabs. Brown, a former Excite@Home and Time Warner Cable engineer, joined CableLabs in 2001 to spearhead high-profile projects such as DOCSIS and CableHome.

CED Editor Jeff Baumgartner recently caught up with Brown to find out how CableLabs plans to cross-pollinate these activities. An edited transcript follows.

CED Editor Jeff Baumgartner recently caught up with Brown to find out how
CableLabs plans to cross-pollinate these activities. An edited transcript follows.


: On paper you are heading up the CableLabs Broadband Access department and serving as the organization's CTO, but give us an idea of exactly what that entails.

Brown: I still have direct responsibilities for the broadband access department, which is DOCSIS and CableHome, so I'm still responsible for specification development, interops and certification for those projects.

With my background (at Excite@Home and Time Warner Cable), I can bring to the job a lot of cross-project, cross-area expertise...the ability to see the data side and take a look at the video side and see how they relate. What happens when you mix a set-top box with a residential gateway? What are the implications from a specifications standpoint? What are the things to think about operationally? For me, a lot of the focus near-term is the integration of these things.

A lot of what came out of NGNA (Next Generation Network Architecture) was integrating services. It's more than just bundling. A lot of the competition will be able to bundle services, whether they do it through marketing agreements or the technology. It's got to be the integration of services on those platforms.

CED: What is the plan to figure out how all of these projects relate to each other and how they talk to one another?

Brown: There are some things that aren't obvious from the outside. We have DOCSIS; we have CableHome; we have PacketCable and OpenCable. Those are fairly discreet projects. They have their own specifications and there isn't a lot of cross-pollination.

Since I took over this job directing the broadband access department a year-and-a-half ago, we've been really trying to foster collaboration across the departments. You will find today inside CableLabs people whose primary focus is CableHome working with the OpenCable folks to talk about how home networking integrates with set-top boxes and some of the standards that should be applied.

I think one of the first set of specifications that really crosses that divide is the DOCSIS Set-Top Gateway (DSG) specification, which is one that I led early on. The [DSG covers] how you embed a DOCSIS cable modem into a set-top box and the additional requirements that are involved.

CED: The DSG spec has been out for some time, so what are the next steps CableLabs is working on? Are you ready to do some interoperability tests now that set-top makers are embedding DOCSIS?

Brown: We originally drafted that specification in 2001, if I remember correctly.

It was a little ahead of its time and there wasn't a lot of uptake. Although it looked like a good target to go after, there were some questions about when we actually need to pull the trigger on that.

In the past year-and-a-half, there's been a great deal of reinvigoration, and it's like, "Now we need to do it." In the first iteration of a spec, if you don't have implementations to validate it, there are always corrections that have to be made. There were some initial efforts led by the Koreans to implement some things. There was a bunch of learning that came back from that.

Now we're getting to the point where people are serious about it; they're implementing it. We've just completed our second interop with DSG (in September). I'd have to say that the DSG interops are probably the most successful interops we've had at CableLabs to date, certainly on the DOCSIS side.

CED: Is that because of the maturity of the technology?

Brown: You are building to DOCSIS, so there is a very mature base you're building on. When we conducted the interops, they were much further along in their development than we would have thought. A lot of it was that the incremental development wasn't that large ultimately in order to add DSG functionality to DOCSIS, so they were able to get there very quickly. Some of the intent of DSG to begin with was to minimize the impact both on the existing CA (conditional access) systems as well as the existing DOCSIS infrastructure. So we tried to design it in a way that was minimally disruptive.

We had 15 manufacturers in our first couple of interops. We're planning one additional interop before cert wave 32, and that's going to be more of a dry run in preparation for cert wave 32, which will be the first wave [in which] we'll be prepared to qualify product for DSG. (Note: cert wave 32 was scheduled to begin in late October.)

CED: Will you be testing both set-tops and CMTSs?

Ralph Brown at CableLabs
Brown: It's a little bit of a lineup for next year when we'll be doing full OpenCable host certification. Here's a case where the products are really starting to tie in together. The new OpenCable Host 2.0 spec references DOCSIS, references eDOCSIS and references DSG. We designed the specs in a way so that they not only cover OpenCable hosts, but cover, to the extent they can, a certain chunk that's proprietary in the embedded CA set-tops.

We'll look at a portion of the box, so we try to carve out that portion in the DOCSIS and DSG specs and eDOCSIS specs to [show] that this is the part that is in accordance with the specification. There's this embedded CA part and video functionality and this sort of stuff that we don't look at; it's up to the MSOs to deal with because it's a proprietary box that they're buying that's different from an OpenCable box where we'd actually look at the CableCARD and the host interface.

Now we can do a process of validation for our members for boxes that they buy, and it's the same general functionality that an OpenCable host would have. The same testing and processes would apply. Furthermore, as an operator, when I go to deploy this stuff, I want to minimize the operations headaches between one and the other–whether it's for embedded CA or removable CA.

CED: What's the connection between CableHome and PacketCable Multimedia?

Brown: In terms of CableHome and PacketCable, there's been a lot of development and focus on PacketCable Multimedia.

With CableHome, the priority and what we care about is QoS. If you're going to deliver video services or multimedia services over a home network you need to have quality of service for that experience to be as good as somebody would watch it on a set-top box or Voice-over-IP on an MTA.

UPnP (Universal Plug and Play), when CableHome first started down this path, didn't have QoS on the radar, so we went out and did our own thing. We're now members of UPnP. In fact, we are on the steering committee of UPnP now, and we're fully engaged in that activity.

So, we solved the home QoS issues. We've got messaging for how to set that up. CableHome specifications cover that. But how do you bridge that back into the access network? It's not just about how do I play a game between one game console in my home with another console in my home, but it's how do I play that on a network, play it outside my home. Or I want to see streaming media come through over the access network instead of playing it from one PC to another. It's not just about the home network. It's about my connectivity to the Internet for services that are out there.

So our big focus for this year and early next year is going to be on bridging access network quality of service with home network quality of service. Once we do that, you begin to see a wealth of opportunity in terms of services that line up. Video and telephony, games, streaming media, high-speed Internet surfing–all those sorts of things–they all play into the home network.

CableHome by itself, or PacketCable Multimedia by itself, brings part of the solution, but together they can aggregate much more powerfully.

CED: Can you update us on how the integration of NGNA is progressing?

Brown: A lot of what NGNA did was trying to drive consensus around what our members are looking for and solidified some requirements going forward. To a large extent we've tried to realign our efforts with those directives. DOCSIS 3.0 is a good example–looking at higher bandwidth through channel bonding. We're looking at a more modular CMTS architecture. What it has done is accelerate some of that work.

On PacketCable, there's a lot of interest in SIP (session initiated protocol) and PacketCable Multimedia and its support of new media types and going beyond the voice application. On the OpenCable side, there's some clarity coming...on the host devices. What are the real requirements on set-tops that [operators] want to purchase?

CED: CableLabs revealed plans for DOCSIS 3.0 earlier this year. Is there going to be a spec this year?

Brown: No. There's a lot of stuff that actually goes into DOCSIS 3.0. There's an architectural component; there's more bandwidth; there's IPv6. You look at IPv6 and ask, "What does that mean?" There's a lot in IPv6, and you need everything in IPv6 to be specified over the access network.

We think that the remainder of this year will be spent on real requirement gathering, real analysis on what are the economics that impact different approaches and distill it down to the technical requirements... then begin the spec development work for DOCSIS 3.0 in earnest in 2005.