We could conference you in via cell phone. Will you be flying the 22nd during the reception or the following day?
Now, while waiting for the party to begin, here's an article on LMDS I missed earlier. Having heard a good report re: MaxLink, and believing there are other Links in NN's future, this seems appropriate:
<<<<
Talk Radio
Prioritizing POTS over LMDS
By Dan Sweeney
A regular contributor to America?s Network
No one?s prioritizing POTS ? neither the local multipoint distribution service (LMDS) infrastructure providers, a swiftly expanding group of manufacturers, nor LMDS license-holders. LMDS broadband (over 1 gigahertz broad) and its proponents are much more likely to discuss services unique to broadband, such as telemedicine, high-resolution videoconferencing, and high-speed Internet access. Plain old telephone service is assumed, and explicitly provided for, in all equipment developed to date and in all service offerings extant today; but promotional literature consistently downplays voice ? almost as if the service itself were a legacy.
Yet, close reading of manufacturer documentation and close questioning of carrier principals tells a different story. While everyone certainly looks to high-bandwidth services to capture the imagination of subscribers, no one doubts that voice will be a staple in any bundled services for the foreseeable future. Ironically, some see LMDS as ultimately more crucial to the future of voice telephony than voice traffic is to the fate of LMDS.
"Fixed wireless multipoint service has the power to free the competitive local exchange carrier [CLEC] from the role of reselling the Bells? services," says Robert Stuart, marketing director at Teligent (Vienna, Va.), the multipoint service provider with the current lead in numbers of subscribers. (Teligent is an operator of millimeter microwave point-to-multipoint networks, but not an LMDS operator in the strictest sense).
"Wireless provides a true alternative infrastructure, something we haven?t had before," Stuart says.
Bernard Herscovich, director of wireless infrastructure engineering at Newbridge Networks (Kanata, Ont.), agrees. "Previously, fiber rings were the only alternative to the local exchange, but they were confined to the business centers in a relatively few metropolitan areas, and the cost of building them has been prohibitively high in many instances," he says. "LMDS allows for comparatively low-cost infrastructure that may be rapidly deployed and constructed incrementally, on a pay-as-you-go basis. It really makes the CLEC a reality."
Let?s assume that, gigabit for gigabit, an LMDS network is cheaper to deploy than rival infrastructure technologies such as broadband satellite, hybrid fiber coax (HFC) and fiber-to-the-curb (FTTC). Will it match the competition in terms of quality of telephone service? Just as important, will it meet the standards upheld by incumbent LECs (ILECs), not only in terms of availability and reliability, but of voice quality, and of priority of voice regardless of other traffic on the network?
Those are the questions confronting manufacturers as they attempt to sell their wares to license-holders. Those are also the questions for license-holders as they attempt to build a subscriber base in the business community.
LMDS Gets a Voice
The LMDS bands were originally envisioned by the FCC as a medium for transmitting subscription television programming, as has been the case for the lower frequency broadband fixed wireless service known as MMDS. A few analysts and equipment manufacturers still allow this brand of POTS as a possible offering within a family of broadband services, but it isn?t in any of the current licensees? announced business plans.
Most have looked at the limited growth in the older MMDS industry ? so-called "wireless cable" ? and concluded that another source of traditional one-way video programming would not be competitive in the current marketplace. Others, including Doug Weiner, director of advanced technologies at Bosch Telecom, cite inherent technical difficulties.
"Twenty-eight gigahertz transmissions don?t propagate well through foliage. If you?re talking about a large residential market, that?s a problem," Weiner says.
It?s also a problem facing wireless carriers attempting to build a competitive local exchange business. But few believe that LMDS license-holders will want to compete at large with entrenched telcos.
Bill Frezza of Wireless Computing Associates (Yardley, Pa.) cites economic considerations that may also keep LMDS licensees away from POTS for the masses. "The equipment is just too expensive," he says. "You have to have a concentration of users sharing a single access point. That makes sense for a single office building or a business park, but providing radios for individual subscribers is not cost-effective. That limits what they can do.
"The LMDS subscriber units are designed to dock with PBXs or Ethernet ports, not with individual devices," Frezza adds.
In effect, LMDS services, unlike traditional local phone services, are to the curb or to the closet rather than to the desktop. As such, they?re more like a shared leased line than traditional circuit-switched POTS. What will this mean in terms of voice telephony services?
"They?ll try to bundle POTS with a range of other services," predicts Adrian Jones, group director of project management at San Jose, Calif.-based Terayon Communications Systems, a cable modem and headend manufacturer serving the competitive cable data and telephony industry. "Essentially, LMDS is in the same position as the cable industry in regard to POTS. The business case for concentrating on a single type of service while building a very expensive infrastructure isn?t good."
Not everyone agrees, however. Indeed, there?s little apparent consensus among infrastructure providers or licensees on any aspect of the business plan.
"We go after customers who have never had a leased data line," says Teligent?s Stuart, appearing to indicate that T1 replacement will be the mainstay of Teligent?s business, with voice something ancillary.
Newbridge Network?s Herscovich objects to this. He says that "T1 replacement might not be a very good idea, with T1 pricing coming down the way it has. We see high-quality voice over IP [Internet Protocol] being a very good market with telemedicine, distance learning, and emergency temporary markets also offering possibilities."
Perhaps the closest to a consensus position is that expressed by Mark Liperator, manager of wireless broadband at Lucent Technologies. "There are a lot of people out there looking for a business plan right now," he says. "We?ve set up a business consulting division to help our customers arrive at that plan."
No one doubts that POTS will play some part in LMDS service offerings. The only issue is the emphasis that the wireless broadband camp will place on it. The most likely scenario is that LMDS carriers will become specialized types of CLECs, pursuing a select business clientele and neglecting, insofar as possible, the unprofitable traffic that must be provided to the masses of residential customers under universal service provision.
In other words, the LMDS carrier as a species appears to be shaping into just the sort of interloper that traditional LECs decry. No one is saying much about the legal implications of this strategy, but because at least some LMDS carriers are offering or plan to offer long distance service as well as a local exchange, legal challenges are possible.
However, since all carriers who have announced plans for buildouts favor an incremental, pay-as-you-go, building-by-building approach, they can always temporize by claiming that they are insufficiently capitalized to offer universal service through their own plants.
The licensees themselves are not indicating much about their strategies in this regard.
So if the approach of offering bundled services to select business customers is apt to be well nigh universal, how big a part of the bundle is POTs likely to be?
"Voice is where the real money is," maintains Suresh Arora, vice president and general manager for broadband wireless who heads Gaithersburg, Md.-based Hughes Network Systems? LMDS equipment engineering program. "Internet access, video streaming and other data networking services only amount to a $3 billion to $5 billion business out of a total $100 billion local exchange market. You can?t neglect voice and expect to thrive, and you can?t afford to do voice poorly. That?s the real challenge facing the industry."
"Data may drive the market for LMDS, but there?s still a lot of money to be made off voice," concurs Volker Tegtmeyer, senior market consultant at the broadband carrier division of Siemens Telecom Networks (Boca Raton, Fla.). "You have to do it. But it has to be of high quality."
Issues of Quality
Quality of service (QoS), in terms of voice telephony, can have more than one definition.
For some, it implies an absence of perceptible latency in voice transmissions. For others, availability is the issue ? uptime vs. downtime across the network. Other measures that may be cited include uncorrected bit error rate. Generally, audibility of compression, latency and availability are deemed the most important.
In wireless broadband systems, QoS relates to the digital transport protocol in use in the system as well as the air link. Few manufacturers have much to say about the radio portion of the system as it pertains to QoS; since point-to-point microwave systems are fairly well-proven in reliability, the extension of broadband to the multipoint environment was apt to pose few problems.
But are point-to-point and point-to-multipoint systems really equivalent?
The standard argument is that a stationary point-to-multipoint system is really a point-to-point system writ large ? at least at the frequencies in question. The basic constraints of power limitations and operating frequencies dictate a cellular network architecture with virtually 100% reuse of spectrum from cell to cell, and considerable reuse from subscriber base station to subscriber base station ? an advantage made possible by virtue of the fact that transmissions take the form of narrow beams to fixed sites rather than the non-directional broadcasts of mobile phone base stations. Governmental mandates further define the nature of the network by conferring upon the licensees a fixed block of spectrum that need not be channelized into any fixed number of sub-bands of predetermined bandwidth; so spectrum can be used differentially from site to site.
Such latitude in the use of the spectrum allows the carrier to dynamically apportion the spectrum according to the service requirements of the individual subscribers. If, for instance, a subscriber wants a fast Ethernet connection between geographically separated work sites within an overall service area, the carrier can set up and tear down such a connection on an as-needed basis so long as the data traffic is fairly intermittent. Within a fat pipe of designated spectrum, the carrier can use asynchronous transfer mode (ATM), IP, or IP or frame relay over ATM to send diverse kinds of traffic simultaneously over a common physical channel.
Nevertheless, Brian Andrew of Triton, a company manufacturing equipment for the 38 GHz band (and an advocate of a ring network architecture over multipoint distribution systems), is prepared to challenge conventional wisdom. Andrew sees the air link effectively subverting the operation of the higher layers within point-to-multipoint systems.
"Because you?re reusing channels within a cellular architecture, you face the same problem as providers of mobile phone service ? namely, power control ? but it?s complicated by the need to overcome rain fade," Andrew explains. "In a millimeter microwave transmission, rain fade is always a problem. You overcome that by boosting transmitter power, maybe by as much as 30dB. But that 30dB of boost represents 30dB of interference in adjacent cells. There?s no getting around it."
Andrew touches on a whole range of issues relating to the air link in LMDS networks, issues which are still being sorted out because no true LMDS network was fully operational in the United States by press time.
(Teligent?s service is a special case, in that the 24GHz frequency band it occupies is channelized, restricting the carrier?s ability to allocate spectrum dynamically.)
There also are questions concerning the nature of the physical plant. Unlike some of the unlicensed networks, these are not self-healing mesh architectures. So the failure of a single hub could take the system out, because transmissions are severely limited in distance. The fact that infrastructure resides on customer premises is also cause for concern.
If equipment suppliers are taciturn regarding the intricacies of the radio portion of the networks, they are voluble regarding the higher levels of the network. And they are frequently at odds with one another on how voice might best be supported within a point-to-multipoint broadband wireless system, and on the extent of the difficulties in implementing voice smoothly across what will almost surely be multimedia networks.
Hughes? Arora sees considerable difficulties for equipment vendors. "Maintaining voice quality will be a challenge, because you?re trying to do everything else at the same time," he says. "ATM is supposed to provide the answer, but there is widespread mistrust of ATM at the fringe of the network. I can?t think of a public carrier who?s using it at the level of the local exchange."
Liperator takes partial exception to this statement. "On the carrier level, ATM in the local loop may not be common," he says, "but we have ATM equipment designed to go on the customer?s premises, and that has voice capabilities. I think there?s a track record for ATM."
Adrian Jones of Terayon (a company that has built its cable TV equipment around an ATM core), says his company is providing POTS over ATM within a wide-ranging pilot program conducted by Rogers Cable of Canada.
Full commercial deployments are scarcely in evidence. Teligent?s geographically limited services within selected metropolitan markets constitute the most extensive use of ATM in a public local area network to date As earlier stated, the Teligent systems are still in the earliest stages of development.
The uncertainties of ATM in the local loop are compounded by the fact that the ATM protocol was designed for the low-loss medium of fiber, not for an air link where bit error rates are typically orders of magnitude worse than for fiber. Aggressive forward-error correction can be added to the system, and retransmission provisions can be implemented, but how well will they work in practice in a fully deployed commercial network? No one really knows. In fact, no one really knows how well ATM functions in any sort of local loop wired or wireless. Yet, nearly all manufacturers of broadband millimeter microwave equipment use ATM at the cores of their networks, though it appears in somewhat different flavors from network to network.
Francoise Vigneron, product manager of broadband wireless for Alcatel (Richardson, Texas), is a firm advocate of bandwidth on demand and shared resources through the agency of ATM.
"We have an ATM switch for our system, and the system as a whole uses TDMA which is based on the idea of bandwidth on demand," he says. "We also utilize special framing techniques for the cells and packets to reduce delay and echo cancellation is, of course, provided as well. The basic technology has been well proven in the point-to-point microwave links we?ve been selling."
"We?re not locked into the traditional wireline paradigm," explains Kevin Curtis, director of wireless media services at NEC (Herndon, Va.), which will be introducing not only LMDS over licensed millimeter microwave frequencies, but also point-to-multipoint ATM-based network equipment operating in unlicensed 5.7 GHz band, and permitting limited mobility.
"Our systems will accommodate the T1 model of circuit voice," Curtis says, "but we can also transmit IP voice which, of course, involves assigning resources in a dynamic fashion and providing bandwidth on demand. Everyone?s talking about emulating the wireline world, but the licensees are going to have to differentiate themselves from the incumbent LECs."
Tegtmeyer speaks of statistical multiplexing, dynamically allocating network resources literally from moment to moment based on the instantaneous requirements of the signals occupying the various channels. "That really allows you to leverage bandwidth."
Will any or all of these flavors of ATM truly provide parity with circuit switched voice over wireline? Certainly, ATM has proven itself in the long distance carrier business as a transport for a mixed traffic of voice and data, and the standard itself was specifically designed to support multimedia while affording QoS levels for each type of traffic. If IP voice can find a market on the local level, then why not ATM, which was designed from the ground up with a voice capability?
"I see a problem with latency," Arora says. "We?ll be employing ATM in our own system, but I?m not sure it?s the final answer."
"ATM prioritization schemes are fine until the network loading reaches a certain point," says Brian Andrew, president of Triton. "Past that point, ATM is not a substitute for dedicated bandwidth. ATM networks are not nonblocking."
ATM transport does include the concept of permanent virtual circuits (PVCs) where bandwidth always will be allocated to a subscriber regardless of whether he is using the system, but PVCs are spectrally inefficient and undercut the rationale for using ATM in the first place ? i.e., a winning combination of spectral efficiency and differentials service levels with guarantees of quality for each.
>>>>>
And one comparing vendors:
americasnetwork.com
Newbridge Networks? ATM switch is perhaps the most flexible, accepting digital set-top boxes for MPEG-1 and MPEG-2 video programming, codecs for videoconferencing, fiber connections and digital subscriber line access multiplexers (DSLAMs). Indeed, Newbridge advocates the use of digital subscriber line (DSL) within the building or the campus as a low-cost means of connecting up a LAN to complement the low-cost metropolitan connection provided by the LMDS network.
Other distinguishing characteristics have to do with the voice capabilities of the systems. NEC?s yet-to-be-released system will support IP voice with guaranteed quality of service. Bosch and Newbridge include Class 5 telephone switches in their networks. The extent to which the various systems will support special calling features has yet to be established, as few manufacturers are providing highly detailed technical specifications.
|
