Advertisement
Articles
Advertisement

DOCSIS 1.1 Sounds All-Aboard Signal

Sun, 09/30/2001 - 8:00pm
Craig Kuhl, Contributing Editor


Advanced spec is a passport to IP services

The much anticipated inaugural run of DOCSIS (Data-Over-Cable Service Interface Specification) 1.1 is about to leave the station, and with it comes a whole new generation of tiered services, revenues and consumers that are expected to climb aboard the 1.1 train.

Fresh from a painstaking 28-month journey through its developmental process at CableLabs and fraught with intense scrutiny by vendors, cable operators and a gaggle of consultants, DOCSIS 1.1, the advanced version of its 1.0 predecessor, is finally ready to roll.

Its destination is a new era of cable modems, set-top boxes, IP-based platforms, tiered services and additional revenue streams, many of which remain undiscovered.

"For the first time, operators will be full service providers and can offer data services they've never been able to offer before, while vendors will probably see the biggest gains in video and voice services. DOCSIS 1.1 is a passport to IP services. It's been a pooled effort and a remarkable accomplishment in only 28 months," says Rouzbeh Yassini, executive consultant to CableLabs and CEO of YAS Broadband Ventures LLC. Yassini not only has been the lead engine in the 1.1 process, but he's largely recognized as the guy who built the tracks, too.

DOCSIS 1.1 sprang from CableLabs' standardization effort that kicked off 1996 with version 1.0, and adds four crucial elements to that standard: Quality of service (QoS); data fragmentation; a security upgrade; and added reverse path throughput.

"The keys are functionality and QoS. They allow operators to create tiered service and guaranteed bandwidth and revenues. There is an immediate customer benefit, and operators can take advantage to get ready for IP services," Yassini says.

The main differences between the two specifications, he explains, are functionality and the cost of network operation. Because of silicon-based integration, one chip now takes the place of three, reducing 1.1 upgrade costs even more. "QoS and fragmentation are combined in the silicon to guarantee bandwidth. That's the key. It serves as a foundation for PacketCable services like voice. It's all linked together, end to end," Yassini notes.

The next stop for DOCSIS 1.1, he adds, is data analysis and recommendations by the CableLabs board of directors for certifiable 1.1 modems and labels for CMTS products. Those are currently underway.

Although most experts agree the full impact of 1.1 is four years away, few doubt the potential upsides of 1.1.

"DOCSIS 1.0 is the Trojan Horse, and 1.1 is the Greek soldiers inside," says Ryan Jones, media and entertainment analyst for the Yankee Group. "That's the progression. With 1.1 you can provision different levels of service, and after that, you look at 1.1 fueling IP telephony. The demand isn't there yet, but now it's easy for operators to defend their choice to move to 1.1. It also allows guaranteed bandwidth, something cable hasn't had in the past and drives both cost-savings and revenues. It's very encouraging," he adds.

Voice-over-IP (VoIP) traffic is encouraging as well, and most experts agree it will drive a growing number of operators to the telephony market, and 1.1 is a big reason. For example, a recent Frost & Sullivan report projects VoIP traffic to grow from $1.39 billion last year to $14.03 billion by 2006, and cable telephony subscribers will pass five million by 2006, according to the Strategis Group.

Yet DOCSIS 1.1's initial impact will be its ability to guarantee speed and bandwidth, experts insist. "The market is becoming more sensitive to dollars. Not everyone wants $40 to $50 price points, so a 256k connection would be available for about $20 a month. Operators can have 'dynamic throttling' of speed and different bandwidth capabilities. And they can charge extra for services like real-time video, interactive gaming across networks and VoIP. It really provides new revenue opportunities," says Keith Kennebeck, analyst for the Strategis Group.

It's those opportunities that are getting the attention of cable operators. "It (1.1) gives us the ability to see traffic and flow, and we couldn't do that before 1.1. We ran trials with PacketCable, and it convinced us that QoS is crucial. We must have 1.1 to compete and can't get it fast enough," maintains Steve Craddock, senior vice president of new media development for Comcast Inc.

Comcast, Craddock adds, might build an entire business unit around its home network services, which are driven by the addition of 1.1 technology, while its business segment is likely to expand considerably with 1.1. "It allows our business segment to handle peaks with other services, so we can manage traffic. It's incredibly flexible and will allow us to deal with service level agreements, offer gaming communities and VoIP services. Customers have been clamoring for defined, guaranteed services, and we can charge for those premium services, and guaranteed modem speeds. There's a whole lot of other things we want to do with 1.1 as well," Craddock says.

Doing them cost efficiently, however, could be an issue. The biggest cost issue, experts maintain, is with the CMTS units, most of which will need upgrading or replacement. "The headend must replace the CMTS and that will cost the MSOs, and they don't want to replace all the modems. It's not a quick flick of the switch, and 1.1 probably won't be deployed immediately by cable operators, who are tied up with getting bandwidth access out there. The economic slump has also shown new light on just how sensitive operators are in deploying new technologies and services," Kennebeck cautions.

Several top MSOs are following that path and deploying a "wait-and-see" attitude toward 1.1. "The technology is ready, and we're watching it closely with a granular approach. We want to be sure we have the back office system updated to support it," says Luisa Murcia, vice president of technology for IP services for AT&T Broadband.

Cox Communications' 1.1 strategy is focused on the business market. "Our first area of attack is the commercial space and small- to mid-size businesses using cable modems for their Internet access. While VoIP is a hot topic, we're watching and waiting and will do some field trials this year. But 1.1 is the platform that makes VoIP a quality, viable product and gives operators the ability to launch IP services in a serious way," says Michael Hale, director of data engineering for Cox Communications Inc.

Cox, he adds, ran a series of QoS and modem interoperability field tests and is accelerating its modem upgrades with pre-approved 1.1 software.

A key piece of information, Hale admits, is just how 1.1 affects Cox's current networks. "How does it change our current Cisco UBR platform and (the ratio of) subscribers-to-chassis? We'll need more CMTSs, so at what point do we stop buying first-generation platforms? That's directly related to cost, and we're not looking at replacing old gear, but how many 1.1 subscribers will it support?"

Hale estimates the cost of additional overhead, added resources and hardware to support 1.1 is 15 percent to 17 percent more than 1.0 gear. "There's sophisticated technology that's different than 1.0," he maintains.

Timing is another issue. Adds Hale: "It will take two quarters before we see the full story of 1.1. But we'll experiment with 1.1 features in select markets to get some real world experience with large transactions and might try out different tiers on live platforms to gather information next year. This is not an overnight thing."

With 1.1, demands on the routing portions of CMTSs will be strained and set-top boxes will be asked to handle more traffic, while software upgrades will be an on-going issue, he maintains. At the end of the day, however, the effort will be worth it. "It's all in the name of future revenues and cost-savings, and we've shifted our focus to adding services and better take rates. That will benefit us in the long run," he says.

In the near-term, however, competition from DSL providers and a fertile small- to mid-size business market are pushing more cable operators into the waiting arms of 1.1. "We're seeing the business market emerge and operators are seeing DSL in their rearview mirrors," says Stuart Bennington, group marketing manager North America for Tellabs Broadband and Media Group. "They recognize that it's time to strike and get market share in the high-speed data market. Operators haven't been aggressive to the business markets; now, 1.1 allows them to mimic private line services of the ILECs with a guaranteed amount of bandwidth. But none of this is a service if you can't bill separately for it, and it's a huge decision to deploy more equipment to the same number of customers," he says.

Huge decision or not, operators and vendors alike are realizing that to ignore 1.1 could be nearly fatal. "We're actively pursuing 1.1 on the modem and telephony sides and have designed our hardware silicon to be 1.1 compatible, which will allow software upgrades to the set-tops for 1.1 capabilities, and 1.1 is the only way operators can get to IP telephony competitive services. Clearly, it will allow all types of configurations with one simple box into the home," says Bill Wall, technical director for subscriber networks for Scientific-Atlanta Inc.

Yet the biggest benefit of 1.1, experts say, is its ability to offer tiered services. "We're watching how 1.1 progresses, but the whole suite of services 1.1 offers and the different levels of service are important to us. It's a technical advancement and a cost analysis must be included, but we're working toward 1.1," says Rich Higgins, director of telecommunications engineering for Time Warner Cable.

So are a host of others. Enthuses Craddock: "We can absolutely define streams, configure services unique to modems or class of service, define priorities and determine needed bandwidth, and we can't do VoIP without 1.1. All of the services it delivers will have an impact on our company's bottom line, and it all gets back to 1.1."

Once the 1.1 train leaves the station, a new era of competitive services is expected to begin. "The key is the platform and turning 1.1 and the whole system into an IP-based network that will allow efficient and merged services. That will be a great platform to work from," concludes Kennebeck.

 

CableLabs lays tracks for DOCSIS 2.0

Aiming to build a more symmetrical relationship between cable's upstream and downstream pipes, CableLabs is also architecting the next generation of its standards-based cable-modem platform: DOCSIS 2.0.

By incorporating advanced physical layer modulation techniques, DOCSIS 2.0 will essentially widen cable's upstream throughput to 30 megabits per second, closely mirroring the 40 Mbps available today in the downstream path. CableLabs also claims the new spec will be backwards compatible with existing DOCSIS 1.0 and 1.1 cable modems and cable modem termination systems.

The next-generation specification will contain two modulation methods: advanced frequency agile time division multiple access (A-TDMA), a standard backed by chip makers such as Broadcom Corp., Conexant Systems Inc., Pacific Broadband Communications and Texas Instruments Inc., and Terayon Communication Systems Inc.'s proprietary synchronous code division multiple access (S-CDMA) technology.

Terayon will provide paperwork behind S-CDMA to CableLabs' royalty-free pool, but will not relinquish its core silicon and implementation, says Terayon Senior Vice President and Chief Scientist Rich Prodan.

"That's the art that we possess. Anyone can build it," he adds.

Including both techniques will give operators a "toolbox" of options for advanced applications such as peer-to-peer networking and video conferencing, says CableLabs Executive Consultant and DOCSIS Project Leader Rouzbeh Yassini.

"It's up to [operators] to decide–not today, but at a later time–what they want to do," he adds.

CableLabs estimates the new spec will be completed this year, followed by certification and qualification testing by mid-2002.

Advertisement

Share This Story

X
You may login with either your assigned username or your e-mail address.
The password field is case sensitive.
Loading