Wireless system designs are built around three primary access methodologies: TDMA, FDMA, and CDMA. These access methods apply to the connection from the customer-premises site to the base station, referred to as the upstream direction. Currently, most system operators and standards activities address the TDMA and FDMA approaches.

In the downstream direction, from base station to customer premises, most companies supply time division multiplexed (TDM) streams either to a specific user site (point-to-point connectivity) or multiple user sites (a point-to-multipoint system design). Figure 4 illustrates an FDMA scheme in which multiple customer sites share the downstream connection. Separate frequency allocations are used from each customer site to the base station.

Figure 4. FDMA Access

Figure 5 illustrates a TDMA scheme in which multiple customer sites share both the downstream and upstream channel.

Figure 5. TDMA Access

With FDMA and TDMA access links, whether downstream or upstream, there are a number of factors that affect their efficiency and usage. For FDMA links, the customer premises site is allocated bandwidth which is either constant over time or which slowly varies over time. For TDMA links, the customer premises is allocated bandwidth designed to respond to data bursts from the customer site. These two access methods will probably provide the majority of access links for LMDS systems over the next few years. The choice between these access links is directly related to the system operator business case, service strategy, and target market.

Large customer premises may require a wireless DS–3 or multiple unstructured DS–1 connections. A customer might purchase the use of this wireless connection with the understanding that the bandwidth is available 24 hours a day. In this case, FDMA access links make sense, because the user is paying for and receiving dedicated bandwidth over the wireless access system as well as over the network infrastructure. Typically, the FDMA links terminate in a dedicated FDMA demodulator circuit within the base station.

The other extreme customer case could be customer premises sites that require a single 10BaseT port for Internet access. These users have very low upstream data requirements (acknowledgment packets and data requests are the primary traffic) but may have fairly large downstream data requirements. In this case, TDMA access makes sense, allowing multiple low–data rate users to share a single channel. In addition, the base station terminates the TDMA access link in a single modem, allowing multiple customers to share the single modem at the base station.

Most system operators will have a service mix and target market that lies between these two cases. The choice of TDMA and/or FDMA access methods within the system becomes an issue both for the system designer and the system operator.

As a final example, suppose a system operator wishes to serve a six-story office building containing 20 employees per floor. This offers a total POTS line count of 120. Each office currently uses various mixtures of frame relay, DS–1, fax lines, and modem lines. Some offices wish to connect their in-office Ethernet local-area network (LAN) to the wide-area network (WAN) using routers. The system operator knows that only a percentage of the offices will switch to a wireless service provider.

How does a system operator decide when to use TDMA and when to use FDMA? First, it is necessary to estimate the peak and average expected traffic data rate from all of the potential or estimated offices. Second, it is important to determine which traffic may be multiplexed and traffic-shaped to smooth out the traffic burstiness. If the resulting burstiness is smooth enough, the upstream traffic requirements can be handled effectively using FDMA techniques. Alternately, if burstiness persists within the traffic stream, TDMA may be a better choice.

There are additional system issues relating to the choice of TDMA and FDMA such as the efficiency of the wireless medium access control (MAC), customer-premises multiplexer efficiency, channel structure efficiency, amount of forward error correction (FEC) used on the channel, maximum data rate during peak hours, sharing of the base-station equipment during commercial off-peak hours, blocking levels allocated to the wireless access links, asymmetrical and symmetrical traffic mixtures, and link distance which can be sustained for the various access methods. These issues are discussed in Table 1.

Issue	TDMA	FDMA
user burstiness efficiency	TDMA allows for bursty response and does not request slots unless necessary.	FDMA link is always on, regardless of whether or not the user sends data.
wireless MAC	MAC efficiency ranges from 65–90 percent or higher depending on the burstiness characteristics of the users and the MAC design.	Efficiency is estimated at 100 percent, no MAC.
customer-premises mix	Both the FDMA and TDMA systems allow higher-priority user traffic to be sent first.	Both systems multiplex various streams through the same wireless pipe.
channel efficiency	Efficiency is estimated at 88 percent, based on preamble and ranging.	Efficiency is 100 percent.
FEC percent	75 to 85 percent	91 percent
maximum data rate	TDMA allows bursting to the maximum rate of the channel, based on fairness algorithms for the wireless MAC and the customer-premises multiplexer.	FDMA provides a constant pipe, with bursting occurring based on fairness algorithms within the customer-premises multiplexer.

Table 1. TDMA and FDMA System Issues