Infrared and Radio Frequency Wireless LANs

To continue our ongoing coverage of wireless networks, in this issue we take a look at infrared and radio frequency-based wireless local area networks (see the November, 1991, issue for an article on mobile metropolitan area networks). There’s no question that wireless LANs represent a key technology in the development of mobile, pen-based or notepad based computing systems. But while there is a lot of excitement and enthusiasm about wireless networks, it is a technology in its infancy (much like pen-based computing) and must overcome a variety of technical and regulatory hurdles before it becomes a major component in today’s computer systems.

First of all, it is not realistic to expect wireless networks to replace unshielded twisted pair (UTP) and other hard-wired networks. Cable offers much higher data transmission speeds than possible with wireless technology and will do so for the foreseeable future. While most wireless networks offer maximum speeds of about 2 megabits per second, UTP systems operate at a minimum of 10 mbits per second and are expected to reach speeds of 100 mbps in the next two years. Fiber optic cable systems (Fiber Distributed Data Interface or FDDI) will achieve even faster transmission rates. In the same time frame, wireless networks will achieve maximum speeds of 10 mbps. Clearly, wireless systems cannot compete in performance with wired networks.

Wireless local area networks differ from conventional wired networks primarily in the physical and data link layers of the OSI Reference Model. The physical layer simply describes the method by which the bits of data are transferred from one node to another. The data link layer (also called Media Access Control or MAC) describes how those data bits are packaged and error-checked. The higher layers generally conform to existing network protocols or use bridges, routers, or gateways to connect to them.

RF and IR

The two methods of implementing the physical layer in a wireless LAN are the use of radio frequencies (RF) and the use of infrared light transmission (IR). Infrared systems are primarily limited to a single room since the receiver and transmitter must be able to “see” each other. There are, however, infrared “building-to-building” networks, in which the transceivers are placed in windows of adjacent buildings.

Infrared has one major advantage over RF — you don’t need to worry about government regulations of the use of radio frequencies. Currently, there are no radio frequency bands for the exclusive use of wireless LANs other than narrow frequency bands licensed by private companies such as Motorola or RAM Mobile Data for commercial applications. The same situation is true in the rest of the world and, to make matters worse, there are currently no international standards for radio frequency data transmissions. Therefore, different countries have different regulations and allocate the same frequency bands for different uses. However, efforts are underway to establish an international standard for RF transmissions.

In the meantime, infrared does not have this problem and is therefore an attractive alternative to RF wireless LANs. The principle of infrared data communication has existed for many years. Back in the late 1970s, Hewlett-Packard introduced its HP-41 calculator which used an infrared transmitter to operate a portable thermal printer. HP, as well as other manufacturers, use infrared communications in calculators and other electronic devices. Of course, infrared is used in remote controllers for TVs, VCRs, and stereos. The same principle applies to the use of infrared communication in a wireless LAN. The infrared light is beamed from one transceiver to another. The light transmission is encoded and decoded at the sending and receiving ends into a protocol compatible with existing network protocols.

A pioneer in the development of infrared networking is Richard Allen, who founded Photonics Corporation (Campbell, CA) in 1985 and developed an infrared transceiver. The first version of this transceiver was developed for end users and directs the infrared beam of light to a passive surface (usually the ceiling) in the room, where it can be picked up by another transceiver. Using a limited passive area for each transceiver, multiple network nodes can be installed in a single room. To set up a room with infrared transceivers, you simply aim the transceivers at a common spot on the ceiling. A green light on the transceiver illuminates when the alignment is correct. To date, Photonics has only developed an AppleTalk/Localtalk version of the transceiver, called Photolink, which operates at 230 kbps. The system has a range of up to 200 meters.

The company has also developed a miniaturized version of the transceiver for use as an OEM product in laptops and notebook computers. This version uses the principle of diffused infrared light, in which the beam of light is diffused throughout the entire area of the room, and picked up by other computers or terminals containing infrared transceivers. Photonics is working on an Ethernet as well as AppleTalk version of the portable transceiver and claims transmission speeds of up to 1 mbps. Prices in OEM quantities for the transceiver start at $20, making them an affordable add-on to pen-based or other portable systems.

While Photonics is currently the main player in the infrared market, we can expect other companies to enter this market. A company called BICC based in Auburn, Mass., has developed a Token Ring based infrared transceiver, which requires a central transmitter to be installed in the room. The IEEE 802.11 wireless networking group is currently working on a standard media access control layer (data link layer) for infrared networks.

RF Wireless LANs

In 1985, the FCC allowed unlicensed operation of devices using less than 1 watt of power in three frequency bands — 902-928 MHz, 2400-2483.5 MHz, and 5725 to 5850 MHz. These frequency bands had previously been limited to licensed industrial, scientific, and medical instruments and are called the ISM bands. Network vendors therefore began developing devices for networking within those frequency bands.

Unlike licensed frequency bands such as those used by ARDIS and RAM Mobile Data, the ISM bands are open to everybody. To minimize interference, the FCC regulation Section 15.247 stipulates that a signal transmission technique called spread spectrum modulation with a maximum transmitter power of one watt must be used within the ISM band. This technique has been used for military applications for many years. The basic idea is to take a conventional narrow band signal and distribute or “spread” its energy over a much wider frequency domain. Thus, the average energy density is far lower in the spread spectrum equivalent than in the original narrow band signal. In military applications, the objective is to reduce the energy density below the ambient noise level so that the signal is undetectable. The idea in commercial radio networks is to be able to receive and send signals with minimum interference.

There are two techniques employed for distributing the conventional narrow band signal into a spread spectrum equivalent. These are direct sequencing and frequency hopping. In the direct sequence method, the incoming bit stream is “multiplied” by a higher frequency signal based upon a pre-determined spreading function. The original data stream can then be recovered at the receiving end by correlating it with the known spreading function. This technique requires a computationally powerful device such as a digital signal processor to correlate the incoming signal.

Frequency hopping is a technique in which the receiving and transmitting devices move synchronously in a predetermined pattern from one frequency to another, hopping at the same time and in the same predetermined sequence. As in the direct sequence method, the data must be reconstructed based on the frequency hopping pattern.

Currently, the fastest spread spectrum device is the NCR WaveLAN spread spectrum radio which uses direct sequencing at a speed of 2 mbps. Current technology allows a range of about 100 meters.

While spread spectrum modulation is a viable method for wireless networking, the current allocation of ISM bands is proving inadequate due to the competition with other devices in these bands. For example, the higher ISM bandwidths (2.4 and 5.8 GHz) have microwave ovens to contend with!

The Data-PCS

The inadequacies of the ISM bands have prompted Apple Computer to petition the FCC to establish a dedicated frequency bandwidth for data transmissions. Called the Data-PCS (data personal communications service), Apple has requested the FCC to allocate 40 MHz within the 1850-1990 MHz bandwidth strictly for the use of data communications. According to the petition, “an allocation of 40 MHz is sufficient to permit several Data-PCS networks operating at rates of up to 10 mbps, for example, to coexist in the same geographic area today, as well as to motivate technological innovations that can lead to higher data rates in the future.”

Apple’s petition states that the company performed extensive testing of the current ISM bands using spread spectrum modulation and found that “there is a strong likelihood of unpredictable, essentially uncontrollable, interference in the ISM bands.” Apple’s testing also indicated that the 1850-1990 MHz frequency band provides the optimum propagation characteristics for data communications.

According to the Apple proposal, the dedicated bandwidth would conform to “Part 15” rules or, in other words, would continue to require spread spectrum devices with a transmitter power of under one watt.

The petition further requests that the FCC require that data be transmitted on the Data-PCS in “packetized form.” This requirement is already met by all major network protocols. In addition, the petition requests that manufacturers should have the right to maintain proprietary encryption schemes for data transmitted via Data-PCS.

The petition requests the FCC to release 10 MHz increments for the Data-PCS in two-year intervals starting this year (1992). So, it will be a slow process before a full-fledged dedicated data radio LAN will be fully functional. The FCC recently announced a plan similar to that proposed by Apple’s Data-PCS, so progress is being made.

In-Building Microwave

In the meantime, there are alternatives to unlicensed spread spectrum through private companies such as Motorola. Motorola recently introduced its first wireless in-building network which supports Ethernet LANs at 10 mbps. Called Altair (not to be confused with the old Altair microcomputer!), the system uses a centrally located control module which can support up to 32 Ethernet devices within an office area of approximately 5000 square feet. The Ethernet devices are connected to “user modules” located on the desktop and transmit data using conventional frequency modulation (FM) at 18GHz to and from the control module.

While the Altair system indeed provides a wireless solution for the desktop office environment, it does not address the need for mobile wireless communications such as provided by spread spectrum radio or by the Photonics portable infrared transceiver.

In the European market, Olivetti has introduced a wireless LAN based on the Digital European Cordless Telecommunication standard (DECT). DECT employs a star topology with each device in the star transmitting to the server at a speed of 1 mbps. The exact frequency range of the DECT LAN has not yet been specified but will be in the microwave frequency range using conventional frequency modulation, probably around 18 GHz. More specific details are expected in early 1992.

Is Security a Problem?

Many proponents of various wireless approaches like to point out that security is a big problem for their competitors. But, according to one engineer we talked to, security is more of a “red herring.” Most, if not all, network vendors in the wireless field use some form of data encryption. In spread spectrum applications, the spread spectrum function itself is a form of data security. And, as the engineer pointed out, hard-wired Ethernet cables make “excellent antennas.” In other words, security is probably not any more of a problem in wireless networks than in other methods of data communication.

Wireless LANs and Pens

As mentioned earlier, wireless networks will be a key component in mobile notebook and pen-based computer systems. In fact, without wireless networking, much of the utility of a pen-based system will be lost. The idea of working in the field with an electronic notepad from which you can query a host database or transmit a fax to the home office is a crucial concept in the pen-based computing model.

From a software perspective, wireless networks require a different approach than wired networks. In particular, network connectivity can be easily interrupted in a wireless network, for example, if the wireless user walks outside of the range of the radio or infrared signal. Therefore, network software for wireless systems must provide methods for seamlessly detaching and reconnecting to the network without crashing either the network or the remote node. Ideally, the network will be able to interrupt a data transmission and pick up where it left off when the network is reconnected.

Go’s PenPoint operating system has addressed this problem with wired networks by incorporating a feature called “detachable networking.” By unplugging the network cable, the PenPoint user automatically suspends network operations. Presumably, the same approach would work for wireless systems. Nevertheless, the question remains how the network server responds to these sudden interruptions. These are issues that must be considered in the development of wireless network environments.

Data-PCS Is a Must

Regardless of the technical challenges that face wireless networking, all will be for naught unless the regulatory challenges are overcome. The FCC must allocate a dedicated frequency band for data communications. As Apple’s proposal points out, the current shared usage of ISM bands is simply too unreliable for data communications. In addition, international standards must also be established, allowing both U.S. and foreign vendors to be able to design internationally supported wireless network systems. Hopefully, the IEEE 802.11 committee and its foreign counterparts will be able to establish working standards in the near future.