The world is moving toward wireless-Internet connectivity at an astonishing rate. But instead of spelling the end of the wired Internet, the acceptance of wireless connectivity has resulted in an impressive collection of hybrid networks. These networks include such technologies as satellite, radio, microwave, digital subscriber line (DSL), cable, and fiber optics.
Satellite communications in particular are becoming increasingly important in connecting the global community to the Internet. In response to this trend, satellite communications will be one of the hot topics covered at this year's Wireless System Design Conference and Expo (www.wsdexpo.com), which runs from February 24-27. Like the conference offerings, this article focuses on the many design challenges faced when implementing a satellite-based solution.
The role of satellite communications is especially vital in remote locations, which are difficult to reach with a traditional wired infrastructure. Now, advancements in satellite-communication technology and the launch of next-generation Ka-band satellites are going to drive down the cost of satellite-based Internet Protocol (IP) broadband systems. As a result, satellite-communication systems will begin to grow in number. Wireless designers must therefore learn to appreciate this technology and its place in the new era of globally connected hybrid networks.
Several projects are currently examining the technical and business challenges faced by satellite-communication systems. For example, the Advanced Internet Satellite Extension Project (AISEP) (www.adec.edu/nsf) provides a clear picture of the technical challenges faced by wireless designers in utilizing satellite systems. Its goal was to develop and deploy advanced Internet services and technologies over satellite infrastructures for educational purposes. This project was funded in part by the National Science Foundation (NSF). It involved the combined efforts of the American Distance Education Consortium (ADEC) and Tachyon, Inc., (www.tachyon.net)—a provider of broadband satellite infrastructure and services. The project successfully developed, tested, and deployed a Tachyon satellite-based system to geographically remote campuses.
Previous-generation communication systems used the satellite for only the forward link from the Web site to the end user. They relied on a landline for the return link from the end user to the network operator. Today's bi-directional forward speeds are typically available from 300-kbps to 2-Mbps services. The FCC restricts the return channel to 256 kbps or less.
Modern satellites provide another advantage over older systems: They use Ku- and Ka-band transponders that automatically respond to an incoming signal. These sophisticated, active transponders receive incoming signals over a range of frequencies. They then retransmit the signals on a different band at the same time. The existing 11-to-17-GHz Ku and the newer 18-to-30-GHz Ka bands are reserved exclusively for satellite communications. By providing wider transmission/receiver bandwidths, they enable greater data capabilities than were possible with older satellite systems.
Several new Ka-band satellite systems will soon be deployed. A North American service launch of Ka-band satellites is scheduled for 2004. It will support the Hughes Network Systems (www.hns.com) SPACEWAY high-speed broadband and multimedia services.
Bi-directional satellite communications, or SATCOM, connect the remote-user site to a high-speed Internet-backbone network (Fig. 1). The satellite link provides a bridge between the Internet-gateway system and the access-point system on the user's 10/100BaseT Ethernet LAN.
The Internet-gateway system contains a large satellite antenna with associated communication electronics. Transmission between the Internet gateway and user access point is accomplished over commercial Ku- and Ka-band satellite transponders. Typically, these transponders are leased from the satellite operator. This broadband connection often provides satisfactory service for typical Internet applications like Web, POP3 e-mail, and FTP.
Many obstacles must be overcome to provide efficient IP-over-satellite links. One of the biggest challenges stems from the poor performance of TCP/IP networks over high-latency or noisy channels. Such channels are common in geostationary satellite links. Latency, or delay, is an expression for the amount of time that it takes for a packet of data to travel from one point to another. Network latency comes from several factors, including actual propagation time, the transmission media, and processing devices like satellite transponders and network routers.
The typical round-trip latency of a satellite link is 0.5 s. This number represents the time necessary for a signal to travel the 74,000 km that lie between the source, the satellite, and the signal's final destination. Though this delay may not seem long, its effect is compounded by the "slow-start" mechanism that's inherent in TCP connections. With slow start, a TCP-based network can determine the channel-throughput connection by starting out slowly and measuring the response time. Succeeding data packets are sent at a slightly faster rate until the speed of the link is known.
To solve this problem, a different link mechanism can be used over the satellite communications. In the Tachyon network, for instance, all TCP connections are terminated and then re-originated at both ends of the satellite link. The satellite connection uses a new protocol that is transparent to TCP/IP and optimized for the satellite environment. Because the Tachyon network controls the satellite link and scheduling capacity, the throughput for TCP connections is known prior to transmission. As a result, the common TCP slow-start mechanism isn't needed.
Another factor that increases latency in IP-based satellite communications is the need for "traffic acknowledgement." The TCP/IP transmission protocol requires extensive handshaking across the connection. This method ensures that the transmitted packets were actually received. Yet this acknowledgment traffic hinders short, high-burst transmissions over high-latency satellite channels.
A solution to this problem is to reduce the frequency of these acknowledgements by improving the bit error rate (BER). A lower BER, in conjunction with a reduced acknowledgement frequency, will greatly reduce the TCP/IP performance due to latency.>
A poor understanding of the effects of atmospheric losses on high-frequency signals can compound data-latency and data-throughput issues. For instance, take the signals in the Ka- and Ku-band. These frequencies perform poorly in heavy rain, during which the BER can climb to nearly 10−8 for a 256-kbps connection. Transmission in clear skies can have a performance level of 10−10 BERs (Fig. 2).
Dave Del Sontro and John Cecil, both Satellite Product Mangers at Agilent Technologies (www.agilent.com), note that the signal-propagation attenuation problem is only one concern arising from the Ka- and Ku-band frequencies. Another issue is the availability of associated high-frequency components, as well as the increased complexity of satellite testing. Such high-frequency signals mean more complex signal interfacing via waveguides, instead of traditional coax cable.
Several recent advancements have made satellite-based technology more attractive to businesses and consumers alike. Phased-array antennas, though not new, provide more efficient bandwidth utilization through frequency reuse. This technique increases communication capacity through the reuse of assigned frequency bands. Satellites can now communicate with a number of ground stations using the same frequency. They accomplish this by transmitting in narrow beams pointed toward each of the receiver stations. The beam widths can be adjusted to cover areas as large as the United States or as small as a single state.
Smaller antenna dishes will be spawned from newer Ka-band satellites. The beam width of the satellite-receiving antenna defines the minimum antenna diameter. The beam width is therefore inversely proportional to the signal wavelength. Compared to Ku receivers, the shorter Ka-receiver wavelength will result in a narrower beam width.
Unfortunately, the Ka-band signal wavelengths—especially around 18 GHz—are so short that rain, snow, or even rain-filled clouds passing overhead can attenuate the incoming signal. This problem occurs because the raindrop's physical length approaches a resonant sub-multiple of the satellite signal's wavelength. The droplet can absorb and depolarize the Ka-band signal.
For regions where rain attenuation will be a problem, Ka-band designers are using larger antennas than would be required under clear-sky conditions. Even a slight increase in antenna aperture size gives the overall system several decibels of margin, allowing the receiver to function during light to moderate rainstorms.
Another significant advancement, according to Dave Sontro and John Cecil, is regenerative satellite-transponder technology. Advanced satellite systems use these transponders to recover the original signal from the modulated carrier. They then process the demodulated signals and remodulate them on the downlink frequency in a different format.
All of these technologies—spot-beam and regenerative transponders—require new DSP technology, power-management mechanisms for limited satellite power, and advanced testing techniques. When combined with the advantages of the Ka-band satellites, these innovations will help broadband satellite operators compete in Internet-accessible hybrid networks.