Network Computingwww.networkcomputing.com

In today's networks, we have a mixture of twenty-first-century demand and 1950s infrastructure capability³¹ (Figure 7-19). The huge capacity of the fiberoptic backbones is part of the twenty-first-century demand. The backbones have grown from the initial few installed by the major common carriers for long-distance traffic to many new miles of fiber not only for long-distance voice but also predominantly for data. In fact, it is the data load from business and the Internet that is driving the tremendous increase in backbone traffic. Wave division multiplexing (WDM) using different colors of light in the fibers came along just in time to rescue the major fiber backbones from being completely out of capacity.

Dataquest (San Jose, CA), in one of its market reports, projected that the number of T1s in the United States will increase from about 2.8 million to more than 5.1 million by the year 2004. With voice traffic growing at about 5 percent per year, most of this increase is driven by the bandwidth needs of the computer age.

However, connecting these two marvels of technology is good old twisted pair. This intersection is where the incompatibilities of the 1950s meet the twenty-first century. Our blazing fast 600-MHz Macintosh computers are connected to the multigiga-bit-per-second fiber backbone with 28.8-kbps modems, 64-kbps ISDN (where available), or 1.5-Mbps T1s—all part of the 1950s technology.

This is a need LMDS can fill easily. LMDS can fill this need and provide high-speed, highly reliable connections from the workstation or LAN to the high-speed backbone. The traffic that LMDS can transport is not only computer data. LMDS also can and will transport voice as well as video.

It is ideal for transporting voice over IP and switched video as well as computer data. This reliable, high-speed technology is best suited to provide the interconnection through the twenty-first century.

Market Characteristics

Understanding the market requirements and needs constitutes the first steps in implementing an LMDS system. The next step is to understand the characteristics of these needs to provide the appropriate implementation of LMDS. To do this, we must examine who the users are and how they use the current technology.

Table 7-2 shows typical characteristics from various business segments.32 On the left are large businesses that generate aggregated data. These are data (voice can be thought of as data even if it is not over IP) that are output from the trunk side of a PBX or the network side of a LAN. For security, most businesses permit access from workstations to the outside world only through a defensive mechanism such as a firewall. All the off-premises traffic is routed through this device. This traffic is relatively constant. PBX trunks are engineered for 90 percent occupancy. The output of a proxy or bastion server is also relatively constant with some burstiness.

If the traffic is steady or bursty with a steady nonzero component, then the best access method is to assign a segment of spectrum to the user—frequency division multiple access (FDMA). If, on the other hand, the traffic is sporadic with periods of no access, then the best access method is one where the spectrum is shared among several users—time division multiple access (TDMA). When one user is not using the spectrum, another one can use it. In effect, the spectrum and the LMDS hub are acting as the aggregators of the traffic³³ (Figure 7-20).

Figure 20 Click here to enlarge

The choice of access method is important—if the users offer steady traffic and you have provided a TDMA access method, there may be no time slots for other users. This results in a blocking situation. For sporadic users, a TDMA system is best. For steady traffic, an FDMA system works best. It is imperative that both types of access methods can coexist on the same LMDS hub.

The mix of users and their traffic patterns on the hub is not known and will change over time. The system must offer and support both access methods so that the hub can be optimized and the most efficient use of spectrum (your scarce resource) can be guaranteed.

Rain Fade

At 28 GHz, rain will decrease significantly the radio energy that the receiver gets. With this, the link must be engineered to achieve the link availability that customers need by including provision, as established by the ITU, for rain fade in the operating region. It is also important for LMDS system design to consider the radio environment. Some of the techniques that can be used are:

• Forward error correction

• Dynamic power adaptation

• Dynamic modulation adaptation³⁴

Transport Protocol

The radio provides the physical link. The choice of transport protocol is very important because it will determine the operational characteristics of the system as well as the types of services that can be offered to the user. The important characteristics are support for all types of cargo (time-sensitive cargo such as real-time video and voice as well as time-insensitive cargo such as e-mail), avoidance of error bursts (system performance must be maintained), and standards that promote availability of low-cost silicon. ATM meets these requirements. It easily supports all types of data and even permits classes of service within the groups of data. All types of legacy traffic as well as currently anticipated traffic can be accommodated readily.

The short (53-byte) packet size enables inexpensive wire-line-speed forward error correction and minimizes the throughput penalty that longer packet sizes impose. An error burst during a 64,000-byte packet transmission will necessitate the retransmission of all 64,000 bytes. With an ATM-based system, retransmission can occur in 53-byte increments.

Modulation

Modulation is the conversion of bits to hertz. Several methods can be used. One of the more popular is quadrature amplitude modulation (QAM). This comes in several flavors—4 QAM, 16 QAM, 64 QAM, and higher. The higher the QAM, the more bits that can be transmitted in a hertz of spectrum. The price for more bits per hertz is the need for a much cleaner signal at the receiving site so that the more tightly packed bits can be recovered. With an FCC limit on the radiated power, this means reduced range. Thus the tradeoff becomes either lots of data-carrying capacity or longer range—which should not be a problem if the LMDS system supports multiple modulation methods in the same sector.

Preliminary market research will show the location of high-speed data sources. Locating the hub within the high QAM range will permit serving customers at the most efficient modulation. Lower modulation methods will serve customers located farther away. This allows the advantages of both to be utilized.

One marketplace certainty is the ever-growing need for bandwidth. With the growth in data, faster processors, video conferencing, data sharing, and distributed offices, the one component that is missing is high-speed access to the desktop. The LMDS marketplace has the spectrum and the technology available today to deliver on this need. The wish here is to never have to see the hourglass symbol pop up on computer screens. Instead, with LMDS technology, the industry can offer instant access to meet all information and telecommunications needs.

Coming Up Next: LMDS/LMCS Obstacles to Growth

21 -    Charles Mason, "LMDS: Fixed Wireless Wave of the Future?" America's Network, Advanstar Communications, 201 Sandpointe Ave., Suite 600, Santa Ana, CA 92707, 2000.

23, 24, 25, 26, 27, 28, 29 -    Lam, Derek, Elrefaie, Aly F., Plouse, Lynn, Chang, and Yee-Hsiang, "Telephony Solution for Local Multi-Point Distribution Service," HP Labs, 1501 Page Mill Road, Palo Alto, CA 94304-1126, 2000.

31, 33, 34 -    Ihor Nakonecznyj, "Marketplace Demand for Bandwidth Here and Now Gives LMDS Edge," America's Network, Advanstar Communications, 201 Sandpointe Ave., Suite 600, Santa Ana, CA 92707, 2000.