Terrestrial network strategists researching the business prospects for evolving satellite capabilities can find reasons for both relief and concern, depending on how far out they look.
For operators of wireline broadband networks, the good news is that, even if Rupert Murdoch's much discussed "Death Star" project clears its regulatory and business hurdles, it won't have the technical wherewithall to offer the range and depth of services that will be available over hybrid fiber/coax and all-fiber networks. Even where retransmission of local TV stations is concerned, capacity limitations of News Corp. 's American Sky Broadcasting system, which will include satellites operated by EchoStar, severely restrict the options that will be available in any given locality.
But this leaves no room for complacency on the part of terrestrial network operators, because the bad news for them is that next-generation Ka-band fixed satellite services (FSS) slated to launch in '99 and beyond promise to alter the parameters for satellite participation in broadband communications beyond the turn of the century. While, in and of themselves, Ka-band systems don't have the capacity to challenge regionally ubiquitous fiber or hybrid fiber/coax networks, they have immense power to augment other competitive thrusts against wireline networks, including those already mounted by DBS, and soon to be mounted by LMDS (local multipoint distribution services).
"There is a lot of potential in a link between satellite and LMDS technology," says Shant Hovnanian, CEO of CellularVision USA, the only commercially operating LMDS company in the U.S. Hovnanian says CV is looking at the possibilities of partnering with various entities, including satellite companies, to obtain licenses when the 1.3-GHz LMDS spectrum block at the 28 and 31 GHz tiers goes up for auction, presumably sometime this summer.
"Satellite systems have the ability to deliver movies and other multimedia entertainment and data to LMDS service areas on a dedicated basis, which could make provision of on-demand type services extremely cost- effective for us," says another player in LMDS, asking not to be named. "When you look at what can be done at Ka-band, you're talking about tight targeting, down to areas a few hundred miles in diameter."
In fact, the target areas may be smaller still in Ka-band. "I've seen plans calling for spot beams covering areas 100 miles in diameter," says Robert Bednarek, senior vice president for engineering and operations at PanAmSat Corp., which is in the process of merging with General Motors Corp. 's Hughes Electronics Corp. "When you get down to those sizes, you're able to focus on metro areas where the service begins to look like a broadcast station service."
Moreover, Ka-band will be a two-way service with uplinks typically running at up to 3 megabits per second, and enough downlink bandwidth available to provide dedicated broadband data feeds (1 megabit per second and higher) to a significant number of simultaneous users within the covered area. What makes these capabilities more potent than most Ka-band business plans would suggest is the maturing of advances that were too new or too proprietary to include in public documents and statements, including, especially, active array antenna technology, which aids in narrowing beam footprints and in maximizing bandwidth efficiency.
All of this is a long way from what Murdoch has in mind, at least in the Ku-band phase of the ASkyB/EchoStar initiative. While much has been made of the spot-beam capabilities of the high-power satellites to be launched by News Corp. starting at the end of this year, the spots are too large and overlap too much with each other to allow the frequency reuse essential to retransmitting the lion's share of local broadcast stations nationwide, notes Abe Peled, CEO of News Digital Systems, a unit of News Corp.'s News International plc based in England.
"Most of the targeting of local channels to subscribers in their service areas will be done through use of the conditional access system," Peled says. "The spot beams are really more important for their role in making more efficient use of satellite power."
Just what the impact of spot footprint overlap will be on frequency reuse, News Corp. officials aren't saying. But the technical parameters will limit the number of stations News Corp. can deliver, largely to major network affiliates, according to Murdoch, which raises regulatory issues and complicates the business of lining up station affiliates.
While the local station component, even with these restrictions, may boost DBS marketing appeal in the battle with cable, the more significant drawback associated with the technical capabilities of ASkyB is that they aren't enough to support delivery of dedicated data services over real-time interactive connections, Peled says. "We prefer the broadcast or push model for delivering Internet services," he notes, describing real-time sessions as "a very inefficient use of satellite capacity."
A very different picture of satellite service capabilities emerges, however, when the Ka-band side of the equation is factored in. Loral Corp.'s plans for its CyberStar service slated to get underway in 1999 come closest to representing a direct assault on the local consumer and small business broadband access market. The company is proposing an all-data service and is looking at co-locating its Ka-band birds with the Ku-band systems it inherited with acquisition of AT&T's Telstar operation, says Terry Hart, president of Loral Skynet, the Telstar operating unit.
"We've had discussions along those lines, including how we might start something like the CyberStar data service in something other than Ka-band and evolve it to Ka-band as the satellites are put in place," Hart says. "The interactive component combined with DBS broadcast could make a very nice fit for satellite technology."
Loral isn't alone in scouting the linkage between Ka-and Ku-band and the possibility of delivering combined packages to hybrid antenna units from co-located Ku- and Ka-band satellites. EchoStar has applied for two Ka-band slots with such possibilities in mind, sources say, and executives at Hughes Communications' Ka-band venture, Spaceway, have indicated they're looking at linkups with DirecTV along similar lines.
Spaceway says its two-way Ka-band transmitter/receiver antennas and supporting components will cost under $1,000, and CyberStar is talking about beating the $500 price point. But, as Bednarek acknowledges, most business plans are either very tentative or still evolving, not only because most launches are still two years or more off in the future, but because the technical cost/benefit equation keeps improving with new advances.
"We're seeing an explosion of possibilities as a result of several key advantages associated with Ka-band satellites," Bednarek says. "The question we and everybody else are asking ourselves is, what are the services and the business models that will generate the best returns on this technology."
Satellite interests are coming at the Ka-band option with varying degrees of enthusiasm, though all see a rising global demand for fixed service capacity amid the latest successes in Ku-band as a strong sign of high value for future capacity. "What you're seeing right now is everybody staking a claim for orbital slots, because they know Ka-band is a tremendous resource," says another satellite company executive, speaking on background. "But, frankly, beyond knowing you've got to do it, there's a good deal of uncertainty here and elsewhere as to how we might make use of this technology."
Most of the fixed satellite service (FSS) applications cited in plans filed with the FCC are pegged to direct two-way communications between 26-inch dishes and geostationary satellites, the exception being the $9-billion FSS low earth orbit system planned by Teledesic Corp. For example, Hughes' Spaceway, with two satellites slated for launch in early '99, says it will deliver 108 Mbps of dedicated services to cells in the U.S. measuring about 400 miles in diameter, which means they'll consume a little over one-tenth the land area of the footprint of a typical Ku-band spot beam.
While there will be 48 such beams covering the whole country from Spaceway's two U.S. satellites, frequency reuse is limited to non-contiguous spots because of the overlap problem, meaning that only one-fourth the total bandwidth is available to any one territory within a spot footprint. This is why the throughput is only 108 Mbps per cell over 500 MHz of total FSS bandwidth.
But much higher ratios of bandwidth to potential users are in the offing. Teledesic, for example, says that, starting in 2002, it will offer the equivalent of 19 simultaneous 1.5 megabit-per-second data links or larger numbers of lower-rate connections at a total throughput of 28.5 Mbps to geographical "cells" with diameters of about 20 miles each representing about 1,250 square miles of territory. This means Teledesic's spot beam payload, one-fourth the size of Spaceway's, will be delivered to a target area that is only one one-hundredth the size of Spaceway's.
Teledesic has inspired widespread skepticism over its founders' plans to spend $9 billion on an infrastructure that uses 840 satellites orbiting 435 miles above the earth to provide two-way broadband services of every description to any point on the globe. But at least Craig McCaw and Bill Gates, with a newly issued license in hand from the FCC that their geostationary rivals lack, now have an opportunity to begin selling potential operations and investment partners on the merits of their plan.
"Until we got the license it was difficult to move ahead, but now we are acting on all fronts," says David Twyver, who recently left a position as president of Nortel 's global wireless operations to become president of Teledesic. "Because the lion's share of the cost won't be needed until 2000, the priority now is on lining up our vendors, contractors and operating partners."
Given the fact that broadband network coverage in the U.S. by 2002 will still be far from complete, high-speed data connections for work-at-home and remote site applications should be an important opportunity for Teledesic, Twyver says. "We definitely see business users as an important market segment for our services in the U.S., but it's harder to forecast the market in developed countries than it is in regions that don't have sophisticated telecommunications infrastructures," he notes.
In justifying costs that are three times the amounts projected for geostationary FSS systems, Twyver says that, along with having superior bandwidth efficiency, Teledesic will offer the only true FSS telephony service, which he says gives it a greater market opportunity than its rivals will have. The claim rests on the fact that, at 22,500 miles, the geostationary link distance imposes transmission delays that fall well below the standard for land-based telephony, whereas this isn't the case at Teledesic's orbital height.
Spaceway officials downplay the advantage. "They make way too much of the small delay as a distinction, since much of their as well as our market opportunities are associated with providing voice services where there are no other means available at reasonable costs," says Wendy Green, spokesperson for Spaceway. "The delay is less than what you often get on transatlantic calls." Indeed, a one-half second roundtrip delay in a voice conversation seems a small price to pay for anywhere-to-anywhere connectivity in a global economy. But the issue is certainly a factor in sizing up Ka-band's competitive role in the domestic communications of countries that have strong terrestrial infrastructures.
This is why, from the perspective of possible geostationary FSS applications in the U.S. consumer marketplace, the primary area of concern to terrestrial broadband operators would appear to be the extent to which these satellites might be used to back up Ku-band broadcast service or support LMDS with delivery of truly interactive data services. The potential rests on how much data can be made available on a dedicated, on-demand basis within a given population area.
One place to look for clues as to what FSS might really offer at a moment when many commercial players won't discuss technical details is the National Aeronautics and Space Administration's Advanced Communications Technology Satellite, now in its fourth year of operation. While ACTS had its problems during the first year-and-a-half of operations, critics have largely stopped challenging its credibility on performance in the wake of successful and ongoing connections involving major corporate and financial interests, as well as government and educational entities.
But ACTS is still widely dismissed by engineers outside the satellite business as a hero experiment that, as one senior cable engineer puts it, "demonstrates you can do almost anything if you have a lot to spend and no expectation of a return on the investment." Indeed, to replicate what ACTS is doing using its techniques would be prohibitively costly, even for the likes of Gates and McCaw.
As described by Louis Ignaczac, chief of the ACTS Experiments office, ACTS is capable of delivering dedicated services in simultaneous data streams across 50 beams from four transponders, using switching and buffering techniques to support what appears to be real-time transmission from the end-user perspective. "We sweep through the beam array at a rate of 1,000 times per second, which allows us to provide continuous data feeds across multiple beams from each transponder," Ignaczac says.
With bursts of data every millisecond briefly buffered at the end user station, there are no discernible delays or breaks in the feed through each beam. Thus, with two beams used for scanning and other purposes, there are 12 fixed beams being served per transponder, as opposed to a single beam per transponder as described in many Ka-band plans.
ACTS uses a cumbersome and extremely costly combination of ferrite switches and waveguides linked to highly-focused horn antennas to achieve these capabilities. But Ka-band strategists can now turn to adaptive array antenna technology that is just now moving into commercial applications on the ground to replicate the type of beam hopping demonstrated by ACTS at what promises to be much lower costs.
In all cases, the basic idea entails use of proprietary algorithms running on digital signal processors to dynamically steer and adapt beams and to eliminate interfering signals. Even NASA, which has to live with the system installed on the ACTS satellite, has found a way to use adaptive array technology by putting such antennas on Air Force jets, allowing them to stay in continuous communication with the satellite as they move across multiple beam paths.
"If we were designing (ACTS) today, we'd use adaptive array technology, which is what a lot of people are doing with Ka-band satellite systems," Ignaczac says.
Adaptive array technology, often seen as an option for future rather than present day applications in wireless communications, is coming on stronger than expected as a tool to be used in mobile and fixed terrestrial, as well as satellite applications. "What you're seeing in our case is a technology transfer from military applications to commercial that rests on 15 years of work," says Richard Minthorne, director of business development for Raytheon E-Systems, which is delivering production models of its adaptive antenna system for testing in cellular markets this summer. "It's not well known technology, but it's not new, unproven technology either."
Performance requirements for antennas in mobile applications are different from fixed link requirements, but the physical properties that make smart antennas useful in mobile can be applied in fixed terrestrial and satellite environments as well, Minthorne notes. "We've chosen mobile cellular as a first application opportunity, but really, the opportunities are wide open across all types of fixed and mobile connections, including satellite," he says.
On the satellite front, the first to publicly say that adaptive array technology is a fundamental part of the plan is Teledesic. "I don't think we'll be the only ones to use adaptive array systems," Twyver says.
It remains to be seen how far adaptive array technology can go in allowing satellites to deliver multibeam targeted communications from a single transponder. But there appears to be every reason to expect that previously assumed ratios of dedicated bandwidth to potential users will be eclipsed when Ka-band satellites go into operation. The best defense on the ground would appear to be to build the customer base for broadband as quickly as possible, which means making sure the networks are in place to support digital interactive services within a very few years.