Re: MSO's Driving DWDM?
Denver and Thread,
Denver this article is right up your alley. Just in case you missed it, I thought I would post it.
Since I'm one of those investors who sold HLIT at 9 right before it's move to 90 (ouch!), it's no wonder this article somewhat surprises me. But I should have listened to Hiram (from HLIT thread) when he said, about a year ago, that DWDM was going to be very important to the MSOs.
In case the article disappears later, I'll paste it below. It mentions these companies by name:
Tele-Communications Inc (now AT&T)
ADC Telecommunications
Scientific-Atlanta
Harmonic Lightwaves
MikeM(From Florida)
___________________________
Cable Operators Begin To Seize The Light
By Fred Dawson, Inter@ctive Week
August 4, 1999 2:52 PM ET
The cable industry has suddenly - and surprisingly - emerged as one of the most potent forces driving Dense Wavelength Division Multiplexing into regional and local networks. Cable operators, with a couple of notable exceptions, have been resisting the use of lasers operating in the 1,550-nanometer lightwave window fundamental to this technology, in part because of concerns that its development would divert equipment makers from improving price and performance of optical products operating at the more common 1,310-nm window.
But now that Dense Wavelength Division Multiplexing (DWDM) is a commonplace feature of long-haul communications, the cable sector, led by AT&T Broadband & Internet Services, is finally warming to the 1,550-nm platform. The reason: It not only offers an efficient way to connect regional distribution points, but it also can serve as a local distribution medium that could revolutionize cable network architectures.
"It's hard to believe, but it does look like the push from cable for use of DWDM on the distribution side is going to be much bigger than it is from the telcos," says Vince Borelli, chairman and chief executiveof Synchronous Group (www. syngroup.com), which pioneered use of 1,550-nm lasers and optical amplifiers in Amplitude Modulation (AM) lightwave systems for the cable industry. "The telcos have shown some interest, but the cable guys are really pushing this now."
Not just for singles
Transmission systems based on 1,550-nm AM, often used with optical amplifiers, are now commonly used by cable operators in single-wavelength optical networks that interconnect master headends with primary distribution hubs. But DWDM and its wavelength-splitting abilities are now entering the picture. Several products on the market can combine eight or 16 wavelengths of Quadrature AM signals in a single fiber.
At the same time, thanks largely to an upgrade design decision by Tele-Communications Inc. (www.tci.com), made last year prior to its acquisition by AT&T, multiwavelength technology is entering the fiber portion of cable's Hybrid Fiber-Coax (HFC) distribution network as well.
"The combination of TCI and AT&T backing this technology has everybody in the industry looking at it," says Randy Schmid, director of marketing for the analog transport systems business unit at ADC Telecommunications (www.adc.com), a leading provider of single-wavelength regional networking systems. ADC introduced an eight-wavelength cable DWDM transport system and other optical products at the cable industry's Western Show in Anaheim, Calif., late last year.
It isn't just older system upgrades that are driving the DWDM phenomenon on the distribution end of the network, says Paul Connolly, vice president of marketing and network architecture at Scientific-Atlanta (www.sciatl.com). "People recognize that with things like video-on-demand, higher levels of data use, cable telephony and high-definition TV, there could be pressure on bandwidth in the most sophisticated systems," he says.
Dedication and More
AT&T Broadband & Internet Services (www.attbis.com) is now using DWDM to distribute dedicated data and voice signals. Eventually, it will use the technology to support video-on-demand from the headend as well, according to Oleh Sniezko, vice president of engineering, who presented AT&T's plan at a conference session during the Na tional Cable Television Association's national show in June.
In the AT&T model, each bundle of data, voice and other nonbroadcast signals destined for the distribution network served by a specific primary hub is put onto a separate wavelength at the headend and then combined with other hub-specific wavelengths onto a single fiber. At each hub, the combined wavelengths are optically amplified, and then one of the wavelengths is optically split out for distribu tion locally, leaving the rest to be passed on to the next hub.
The dedicated wavelength that's been split out from the DWDM package is then fed into each fiberserving each node. Added to that feed is a 1,310-nm feed of broadcast video channels that come into the hub over a separate fiber from the headend. The two wavelength signals operate at different power levels, thereby avoiding any interference.
This distribution topology eliminates any need for regenerating signals at the hubs, says Mark Trail, director of product line management for transmission systems at Harmonic Lightwaves (www.harmonic-lightwaves.com), the first vendor to begin supplying the DWDM system to AT&T.
"We have several other large cable customers who are making use of this technology," Trail says, noting that most are applying the technique in single systems before going to wider deployment.
One Way for Now
DWDM also has entered the competition as a possible return-path solution, although, as Sniezko noted in his presentation, costs are not yet competitive and options are less appealing.
"Unfortunately, our bosses did not give us more time in their aggressive plans for advanced service deployments," he said. "Hence, we have to use other tools that are available to us today."
The cable industry needs to find new ways to deliver return signals over fiber from the node, so that it can segment the coaxial portion of the node service area into separate paths. Each path then becomes a dedicated signal that must be multiplexed with the other coax return signals for delivery over the fiber back to the hub.
Segmentation of the coaxial return paths improves operations and maintenance by reducing the number of amplifiers that must be cascaded between the node and any given subscriber. Segmentation also achieves higher bandwidth efficiencies over the 5-megahertz to 50-MHz region of the coaxial return spectrum, because fewer users are connected to any one return path.
In lieu of using DWDM in the return from the node, AT&T has found that the return paths can be effectively multiplexed together in two different ways: frequency stacking, which involves assigning each path a separate frequency and then inserting all the frequency segments into the laser simultaneously; and converting return signals to digital baseband and then using conventional Time Division Multiplexing (TDM) to combine them. Either of these methods will yield a cost savings of 15 percent over a pure DWDM model, according to Sniezko.
Vendors are just beginning to address the TDM baseband option. This summer, Scientific-Atlanta is introducing low-cost TDM baseband return transmitters that are miniaturized for installation in environmentally hardened node modules, Connolly says. In Scientific-Atlanta's scenario, signals from separate nodes would be stacked together at the hub using TDM and then sent out over 1,550-nm lasers onto a DWDM pipe carrying multiple wavelengths of TDM signals from other hubs back to the headend.
Active with Passive
This combining of TDM signals and DWDM technology hints at what could become a revolutionary transformation in the overall architecture of cable networks, affecting not only the bandwidth efficiency and operations performance but also leading to new types of modems and other terminal components. Here again, AT&T is leading the way, with a major test of an advanced system, dubbed Multiplexed Fiber Passive Coaxial (MFPC), now being prepared in Salt Lake City.
"This technology certainly has merits, in terms of an increase in channel capacity by a factor of almost 10," said Xiaolin Lu, district manager at AT&T Labs, at the June presentation. "It provides lower power consumption over digital paths and can simplify the terminal and access protocols we use."
The MFPC design calls for a two-phased implementation that starts with extending fiber deep enough in the network to create an all-passive coaxial path (a path with no amplifiers beyond the fiber node) to each subscriber. "The only new fiber deployed is the fiber around the last couple thousand feet of the coaxial branches, while the existing fiber from the fiber node location back to the hub will be kept and, depending on the design, could even be reduced," Lu said.
Placement of the mini-node fiber termination points can be set up so that only one mini-node is needed for every two and a half amplifiers, according to Lu. Rather than 750 or more households accessing the 750 MHz of bandwidth carrying broadcast and dedicated signals, as is the case in current upgraded two-way systems, the mini-node architecture will reduce the number of homes competing for bandwidth to between 50 and 100, Lu said.
All the fibers to the mini-nodes are connected to a multiplexing hub, which could be located at the fiber termination node or used to aggregate several nodes. The delivery of dedicated signals to the mini-nodes relies on DWDM in an extension of the model AT&T is already using, with TDM envisioned for multiplexing multiple return feeds together at the mux node for the return path to the hub.
Operational Savings
In addition to lowering subscriber counts on bandwidth streams, phase one of AT&T's MFPC implementation will reduce maintenance costs and lower power consumption resulting from elimination of the sensitive amplifiers.
While the current cost of creating this architecture comes to an additional $40 per home passed, the operational savings modeled so far by AT&T total about $11 per year, Lu estimated.
But it's the second phase of implementation that represents revolutionary change.
In phase two, the means of delivering signals would be altered to fully exploit the full carrying capacity of the passive coaxial line, which is in excess of 1 gigahertz, in conjunction with converting to TDM-based transmissions of Internet Protocol data, voice and other services.
"At phase two, we can deliver purely packet-based services at baseband all the way to the mux node, demultiplex those signals and send them to the fiber mini-node, and then transmit them to the customer," Lu said. "We can do similar things in the opposite direction."
By adding a high-frequency tier of TDM-based packet signals to the broadcast analog and digital video signals, while operating the 5-MHz to 50-MHz return channel in the TDM baseband mode over the fiber return segments, AT&T will be able to greatly simplify the media access control and modulation requirements at the end-points of the network, Lu said.
Rather than having a complex stack of gear controlling the allocation of bandwidth parameters and quality of service to each user from the headend, which requires the expensive types of end-user modems in use today, this new system would deploy simple off-the-shelf Ethernet or other standard protocol chip sets at the customer premises. In effect, the technology turns each distribution network into a simple local area network, Lu said.
Testing The Waters
This is where the cost benefits shift overwhelmingly in favor of the new architecture, making the prospects for delivering packetized voice and other advanced services far cheaper than originally projected using current HFC architectures. AT&T is building a 520-mile segment of the MPFC architecture passing 66,000 households in Salt Lake City to test this hypothesis, with plans to begin delivering services in October, Lu said.
"We'd like to work with vendors in terms of determining the technical feasibility and to verify the costs and time to market with this infrastructure," Lu said. "All the data will be collected by the end of this year."
At that point, AT&T will decide whether to progress to phase two of the plan, according to Lu.
Whether or not the AT&T strategy prevails, it's clear that cable network engineers will be putting DWDM to ever greater use as they consider ways to upgrade their networks. The new consensus in the cable industry is that sticking with the tried-and-true optical techniques of the past is not a prescription for success in the future.
zdnet.com |
