x
Cable modem/CMTS

M-CMTS Is All in the Timing

Symmetricom Inc. (Nasdaq: SYMM), maker of a timing element key to the emerging modular cable modem termination system (M-CMTS), is having trouble predicting when cable operators plan to shift gears and adopt the new platform.

Although Symmetricom remains "on track" with field and lab trials with operators, "we cannot forecast the pace at which MSOs will cut over to modular CMTS equipment," company president and CEO Thomas Steipp said Thursday during a conference call about the company's fiscal fourth-quarter results. (See Symmetricom Reports Q4.)

Symmetricom, he said, has started shipping its new cable product, the Time Creator 1000. For the year, those shipments have generated about $900,000 in revenues, with most of that activity happening in the last two quarters.

The modular version of the CMTS is designed to separate out functions -- including upstream and downstream capacity -- that are typically tied together in the same CMTS chassis. This way, operators can scale downstreams and upstreams based on demand.

To keep the clocks of those separated elements synchronized, CableLabs has developed a Docsis Timing Interface Server (DTIS) spec. Last August, Symmetricom's Time Creator 1000 became the first element of the M-CMTS architecture to win approval from CableLabs. (See CableLabs OKs First Modular CMTS Element .)

While operators may use the M-CMTS to support Docsis 3.0, a spec that uses channel bonding to produce shared speeds of more than 100 Mbit/s, it is not required. CMTS vendors are also supporting Docsis 3.0 in integrated chassis.

Few MSOs have yet deployed, or announced big plans to deploy, the M-CMTS architecture. However, vendors such as BigBand Networks Inc. (Nasdaq: BBND) have disclosed some limited M-CMTS deployment activity, most of it happening overseas. (See M-CMTS: Turning Japanese and BigBand Goes Dutch With M-CMTS.)

Among U.S. MSOs, Comcast Corp. (Nasdaq: CMCSA, CMCSK) has reportedly questioned the current business case for making the move to the M-CMTS. Comcast instead will leverage its extant base of integrated CMTSs to embark on initial Docsis 3.0 trials. (See Comcast Preps Docsis 3.0 Trials.)

— Jeff Baumgartner, Site Editor, Cable Digital News

alchemy 12/5/2012 | 3:04:08 PM
re: M-CMTS Is All in the Timing From where I sit, the M-CMTS is a total waste of effort that won't be widely deployed. The problem statement 3 years ago:

* Prices for downstream RF/QAM in the carrier class CMTSs were too high
* Upstream and downstream were on the same blade so it was hard to adjust and balance the mix of US and DS traffic as P2P and VoIP became widely deployed
* Cisco was getting their clock cleaned by Arris since Cisco didn't have a true carrier class DOCSIS 2.0 platform with near-hitless redundancy features and live upgrade

The Cisco bigots at Comcast (now either working for Cisco or ejected from Comcast) drove a NGNA to leapfrog the superior implementations from Arris, Motorola, and Bigband and had made some progress on M-CMTS before launching the initiative at CableLabs. A typical Cisco defense tactic when losing market share.

The problem with M-CMTS (besides complexity), is that redundancy doesn't work very well. What was really needed was to split upstream and downstream into different blades on the CMTS, increase density, and lower the price while preserving the good redundancy story for mission critical VoIP and emerging IP video services. The 3 non-Cisco CMTS vendors went out and did just that and were also forced to squander valuable R&D budget on M-CMTS even though it is a bad idea.

DOCSIS 3.0 is also somewhat nonsense. What was really needed was downsteam channel bonding and IPv6 that was in the proposed DOCSIS 2.0b standard. That could be built with existing silicon and wouldn't increase the cost of the cable modem significantly. Somehow, I don't believe full DOCSIS 3.0 cable modems will be widely deployed since they cost significantly more. The vendors already have lousy margins so this will be passed on to the MSOs. The silicon vendors get hit with huge NRE for features that are too expensive to bother upgrading all the cable modems and CMTSs. The argument is 'commercial' but those customers will likely be served by their own fiber rather than DOCSIS. You don't want a super-fast upsteam since it floods your network with BitTorrent and other P2P traffic. Upsteam channel bonding also has big issues with jitter, delay, and redundancy for VoIP and other mission critical applications. I think it's a bad idea.
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE