Multimedia on tapJan 1, 2003
The satellite systems we have now are largely designed for providing voice connections. The data capabilities of these systems are typically limited and expensive. What if satellite systems were to come on line with true broadband data capability? The effect of this on voyagers could be revolutionary: You could request voyaging information; it would be placed into multimedia data streams and delivered to your boat in real time. (For an idea of how this type of broadband capability might work on a voyage, see the accompanying sidebar.)
Like many technology changes, the driving force for these systems is going to come from government and commercial interests that want the same level of interactive connectivity with their distributed interests that they have with their fixed ones. The trickle-down to the voyager is going to take place at a fast rate because of the incredible level of computational services available on board a voyaging boat and the lower prices for equipment due to a huge demand for "home" and mobile personal communications technology.
And it's going to change navigation in ways Bowditch, Cook, and Maury would have recognized as valuable.
These technologies and concepts are being proven on a daily basis by land-bound users. Shore-based providers are already serving the coastal navigator's need by broadcasting from towers near harbors in much the same way a PCS phone system or pager service would, but data rates are low.
For ocean navigation satellite broadband, there are challenges to overcome, however. One is that ocean users are truly at the edge of the market; another is the need for receiving equipment with a small footprint that can handle a sailboat moving around wildly in a seaway. But there is so much working in favor of this evolution it isn't a matter of if, it's a matter of when.
There is no doubt: wire and fiber alone cannot satisfy the anticipated global bandwidth demand. Satellite broadband systems will be built, perhaps as many as five providers will place $30+ billion of hardware into service in the next seven years. The issue for voyagers is whether their needs will have any impact on service delivery or whether they'll have to learn to adapt to essentially shore-focused services.
Emerging low-earth-orbit (LEO) and medium-earth-orbit (MEO) systems are built around alternate architectures that improve broadband performance for mobile users, provide high-service quality with low levels of latency (signal delay due to the length of the propagation path), and small terminal size. These systems will offer a service price roughly equivalent to what is being paid today for wire-line services to home users.
This is a case for which the technology must precede market demand - user "pull" is not great at this point, and business risk is high. To get those risks down to the investment level, progress on three technology fronts is required.
1. Delivery technology, which sets the fundamental technical parameters for service, including the design of the satellite constellation, the signal processing schemes, and frequency to be used by the system.
2. Service technologies, which create the environments defining how end-users interact over the system: such items as latency, quality of service, and the software protocols used.
3. Applications technology, which recognizes the receive-end limitations. This is the smallest challenge since wire/fiber-based internet applications are already well developed.
Constellations and processing: LEO's low altitudes permit fairly low-powered transmitters for both up and down data transfer. Latency is small for the same reason - both real pluses for mobile Internet operations. But fast-moving LEO satellites can't see much and can't see it for very long, so lots of them are needed if worldwide, uninterrupted operations are required. These swarms of satellites must either be able to hand traffic off to one another independently (like Iridium's satellites) or have a robust ground-based system directing signal traffic.
Frequency selection: The difference between the highest and lowest frequencies of a frequency band is referred to as "bandwidth." Multimedia requires high frequencies to generate the larger bandwidth required to pass data fast enough to satisfy the customer. With careful processing, error correction, and so on, a single 2 megabits-per-second (Mbps) data stream needs about 2.7 MHz of bandwidth.Satellite communications frequencies are generally between 1.6 to 30 GHz, and to prevent signal conflicts, frequency bands used for satellite communications typically have bandwidths of 250 MHz. The lowest bands truly practical for multimedia, L-band and S-band, are used by the big LEO systems for telephone and paging, short e-mail service, and satellite control functions. Next up, C-band is a heavily subscribed feeder link. These frequencies are simply unavailable.
Satellite broadcasting on Ku-band (10 to 18 GHz) is being used for satellite Internet service now and is planned for future systems. Some concepts use Ku up and down, but for interference reasons, this requires spatial beam separation - a potential problem on small craft with small antennas. A few companies plan to go Ku down and Ka up.
Ka-band (18 to 31 GHz) can support 19.7 to 21.2 GHz (down) and 29.5 to 31 GHz (up) for multimedia. However, antennas at this frequency can't be as small as this short wavelength would normally allow due to problems with signal loss in heavy rain. At this writing, 20 license-applied-for program concepts employ Ka-Band.
V-band (40 to 75 GHz) offers much wider bandwidth, and there are plans to use frequencies around 40 to 50 GHz. At least, the license applications have been filed. The technology is a bit of a challenge at this point. It's likely this band will find greater use among various schemes to use high-altitude robot aircraft that will fly above a certain area, providing local and regional relay systems. These could be useful to coastal and blue-water navigators alike.
Managing latency: Even at the speed of light, the shorter the distance the better, and this issue becomes exponentially more important as the complexity of the data to be passed increases. So, geosynchronous satellites will work fine for asynchronous packet-data services, such as e-mail. However, any plan to use packet data networks to support synchronous communications, such as voice and video broadcast service, will be looking for the short hops provided by LEO satellites and the greatest nearby caching (big memory) possible.
Quality of service: In addition to latency, the quality of all services (asynchronous or synchronous) may be degraded by rain interference and other power vagaries. But it's not all bad news. Earth-bound Internet services often depend on interconnected networks to deliver transmissions. This means high-quality carriers may see their data degraded if it must travel over a low-quality carrier's network to complete the hops. Satellite systems can provide total end-to-end connectivity over a single network with quality of service without interconnect degradation.
However, the quality of service benefit hierarchy is still going to be fixed terminals, nomadic terminals, and mobile terminals, due to the fact that mobile users have the least effective antennas. But quality is fitness for use, and as long as the navigator is getting multimedia fit for the intended use, then the issue of quality is moot.
Protocols: Transmission control protocol/Internet protocol (TCP/IP), digital video broadcasting-satellite (DVB-S), and asynchronous transfer mode (ATM) are the entry-point choices for protocols. TCP/IP protocol wasn't designed for satellite links and it doesn't handle even mild latency problems all that well. With the TCP/IP protocol, messages are transmitted in packet form, where each packet may reach the recipient via a different route, arriving at a different time to be integrated at the receiver. While work to correct this problem is underway, this work isn't really needed by very-low-latency earth-bound systems, and TCP/IP, degraded by a lot of scheduling overhead code on each packet, isn't where the satellite community wants to go.
DVB-S is setting new standards for satellite multimedia broadcasting. Among many other things, the protocol addresses effectively encoding MPEG-2 (motion picture experts group) signals for smooth and seamless delivery.
ATM is getting significant attention from satellite systems designers. The ATM protocol transmits data that has been placed in 53-byte, constant-length groups. ATM guarantees data transmission at a rate ranging between 2 Mbps and 2.4 Gbps. The protocol sets up a "virtual channel" between two points based on need and orders the data arrival. The ATM protocol still allows data transmission along diverse paths, though.
Data rate: The LEO systems under discussion are headed for 64 Mbps downlink and 2 Mbps uplink. This is good news for small-craft navigators who will be able to trade data rate for poorer antenna performance and still get the equivalent of seven to 30 times the speed of current dial-up services.
Standardization: The good thing about the Internet is it isn't over-standardized; it can grow in any direction that raises effectiveness, and ineffective modes aren't kept alive by irrelevant standards. The bad thing about the Internet is getting messages like this from your Web browser: "to view this page you must download ..." or, "your browser doesn't support ..." However, when it comes to critical failure-mode issues faced by navigators, some level of standardization seems appropriate
Applications: Fundamentally anything you are using on a laptop today on an earthbound network is going to work. In the near future, at the laptop end, it will all be about wireless interconnectivity on board and graphics, graphics, graphics. From the Internet side it is going to be about forward caching, virtual presence (that doctor is going to seem to be there) and value-added synthesis of products from multiple independent feed streams.
Price: Once government and commercial users have picked up the market entry costs and covered the overhead through bulk use, it is reasonable to expect the marginal cost of these services to price in at close to what household DSL does today (adjusted for inflation).
Capital formation: Cost overruns during development are almost the rule when deploying new space systems. Launch failures, unexpected development costs for user equipment, costs for offshore ground station development and construction, and delays due to government intervention have plagued nearly every space venture since 1960. When governments were footing or subsidizing the bill, this was sort of a "so what?" Not so today.
All of the proposed systems are after the same financial resources of big institutional investors and the public capital markets everyone else is trying to tap into. On top of that, the broadband satellite industry's self-announced growth is heavily reliant on very rosy forecasts of demand for bandwidth across the world, and there are competing technologies out there.
To provide services at the wholesale level, and to satisfy some end-users, satellite networks will have to connect to earth-bound networks, and frankly not all those networks are necessarily outsider friendly. The Federal Trade Commission and FCC both have been asked to look into this when recent telecom-entertainment merger proposals have been under review.
Obtaining licenses from many countries and development of local distributor networks is going to make problems with capital formation look puny. Broadband satellite systems will compete directly with landline and terrestrial wireless communications carriers, many of which are government-supported or are owned and operated by governments, or by very well-connected "friends of the family." The farther voyagers extend their reach, the more satellite services may be the only broadband services.
What's going to make this work is the farmer who wants/needs multimedia to his GPS-equipped, air-conditioned mobile agricultural service unit, or the petroleum geologist trying to squeeze the last drop of oil out of the planet while in the middle of a desert somewhere.
While boat motion is a problem, it is a challenge already met by current satellite systems. And the production of more small, stabilized antenna systems will mean lower prices. What's really going to be a problem is that a lot of the system architectures out there call for the satellites to go to "sleep mode" when over ocean areas to save on battery life. Because the batteries aren't being worked as hard, this sleep mode also allows for smaller solar cell arrays and a shorter battery charging cycle.
A sobering note
Unfortunately, the track record of market assessment and forecasting in the satellite industry is mixed at best. The widely anticipated delivery of anytime, anyplace telephony from systems such as Iridium, to use a rocket-science term, "flopped." Only government intervention has saved Iridium.
Broadband satellite systems require billions of dollars for satellite construction, ground station networks, launch costs, and other expenses. LEO and MEO systems must have enough satellites in orbit to provide uninterrupted global coverage before they can enter service - a big up-front investment. Current estimates show that the multiple systems being actively marketed today - Spaceway, Teledesic, etc. - will require $30 billion, and this number could increase by $100 billion if all the other concepts filing for licenses go to construction.
Allied Business Intelligence estimates revenues from satellite broadband in the U.S. will rise from $25 million this year to $21 billion in 2007. Andersen Consulting pegged the numbers at $8 billion in 2002. Pioneer Consulting has the numbers at $11 billion in 2003 with growth to $30 billion in 2005 and $37 billion by 2008.
With this kind of market potential (approximately 11% of the total broadband market), these systems will come to pass. The return on investment is there. The main challenge will be to ensure that the voyaging market gets good service rather than merely an afterthought.