Standards & Protocols

Fragmentation

The wireless medium on which the 802.11 WLANs operate is different from wired media in many ways. One of those differences is the presence of interference in unlicensed frequency bands, which can impact communications between WLAN NICs. Interference on the wireless medium can result in packet loss, which causes the network to suffer in terms of throughput performance.

Current 2.4-GHz 802.11b radios handle interference well because they support a feature in the MAC layer known as fragmentation. In fragmentation, data frames are broken into smaller frames in an attempt to increase the probability of delivering packets without errors induced by the interferer.

When a frame is fragmented, the sequence control field in the MAC header indicates placement of the individual fragments and whether the current fragment is the last in the sequence. When frames are fragmented into request-to-send (RTS), clear-to-send (CTS), and acknowledge (ACK), control frames are used to manage the data transmission. Therefore, using fragmentation, designers can avoid interference problems in their WLAN designs.

But interference is not the only headache for today's WLAN designers. Security issues are also a major concern. To solve potential security problems, the IEEE has incorporated a MAC-level privacy mechanism within the 802.11 specification, which protects the content of data frames going over a wireless medium from eavesdroppers. The mechanism, dubbed wired equivalent privacy (WEP), is an encryption engine that takes the contents of the entire data frame and passes them through an encryption algorithm. The encrypted data frames are transmitted with the WEP bit set in the frame control field of the MAC header. The received encrypted data frames are decrypted using the same encryption algorithm employed by the sending unit.

The encryption algorithm used by 802.11 is RC4. RC4 is a symmetric stream cipher developed by RSA Data Security, Inc. that supports variable key lengths up to 256 bits. The standard specifies a 40-bit key, but many 802.11-compliant products shipping today support key lengths of up to 128 bits.

Scanning

In order for a mobile station to communicate with other mobile wireless NICs in a given service area, it must first locate those wireless NICs or access points. To enable communication between the mobile station and the NIC, active and passive scanning techniques are supported in the MAC.

Passive scanning involves listening for traffic only on an 802.11 network. Passive scanning allows a mobile wireless NIC to find an IEEE 802.11 network while minimizing DC power consumption. In this mode, the wireless NIC listens for special frames called beacons and probe responses, while extracting information about the particular frequency channel. Although passive scanning expends minimal power, the cost is the time spent listening for a frame on a channel that is idle or may never occur.

Active scanning, on the other hand, requires the scanning wireless NIC to transmit and receive responses from 802.11 wireless NICs and access points. Active scanning allows the mobile wireless NIC to interact with another wireless NIC or access point. The 802.11 standard does not specify a method for scanning. However, many WLAN OEMs support both methods and variants to differentiate their products in the market.

Moving to higher frequencies

When developing WLAN systems, choosing the right modulation and frequency band should be a priority in RF design, especially when designing IEEE 802.11a radios.

For the past decade, WLAN systems have been designed to operate in the unlicensed 2.4-GHz frequency band. The 2.4-GHz band provides 83 MHz of total contiguous bandwidth, spanning from 2.4 to 2.483 GHz.

Moving to the 5-GHz band offers over three times the operating bandwidth over the available spectrum in the 2.4-GHz band. The 5-GHz band is also less susceptible to interference, unlike the 2.4-GHz unlicensed band, which shares spectrum with other wireless appliances such as Bluetooth devices.

There are, however, a few things to consider when switching to 5 GHz. The first is that the frequency allocation isn't contiguous across the band, and the transmit power levels are restricted depending on which block of frequency is occupied. Secondly, in order to achieve the same effective range as covered in the 2.4-GHz band, the transmit power of a 5-GHz system must be slightly increased. As a designer of 5-GHz radios, these issues must be carefully considered in product development.

In the US, 300 MHz of bandwidth is allocated in the 5-GHz band to WLANs under the rules of the Unlicensed-National Information Infrastructure (U-NII). The bandwidth is fragmented into two blocks that are noncontiguous across the 5-GHz band.

In Europe, only Hiperlan WLANs are allowed to operate in the 5-GHz frequency band. A total of 455 MHz of spectrum is allocated for Hiperlan radios. The frequency spectrum allocations for each of the geographic regions are shown in Figure 1.

It's important to point out that although the PHY specifications for IEEE 802.11a are similar to the Hiperlan2, radios compliant with the 802.11a specification are not allowed to operate in the 5-GHz band according to ETSI rules. Efforts are under way by IEEE 802 and ESTI together with the ITU-R to harmonize a global allocation of 5-GHz spectrum for WLANs. Global harmonization could occur by late 2001.

A common approach

When the IEEE 802.11 began evaluating proposals for 802.11a, the working group adopted a joint proposal from NTT and Lucent that recommended orthogonal frequency division multiplexing (OFDM) as the baseline technology for 5-GHz WLAN systems. OFDM was chosen because of its superior performance in combating multipath. This battle is extremely important, particularly in applications that transmit streaming video.

During the development of the 802.11a specification, ESTI was charging ahead with a 5-GHz WLAN project called Hiperlan2. They too adopted OFDM.

For the most part, the PHY for 802.11a is similar to Hiperlan2. The differences between the two standards are minimal and reside in the method by which convolution encoding is used to generate the OFDM symbols and data rates. But it has been said that by making the convolution encoder a programmable feature in the baseband processor, the same silicon can be used to support both standards. This is an extremely attractive feature for those who want to develop products for both standards. Unfortunately, the MAC layers are very different. But that's a subject for a future column.<

5-Ghz channels

In the U-NII band, eight carriers are spaced across 200 MHz in the lower spectrum (5.150 - 5.350 GHz) and four carriers are spaced across 100 MHz in the upper spectrum (5.725 - 5.825 GHz). The channels are spaced 20 MHz apart, which allows for high bit rates per channel. The channel scheme used for 5 GHz is illustrated in Figure 2.

As Figure 2 illustrates, there are 52 subcarriers per channel in the 5-GHz band. Of these channels, only 48 carry actual data. The remaining four subcarriers are used as pilot tones, which assist in phase tracking for coherent demodulation. The duration of the guard interval is equal to 800 ns, which provides excellent performance on channels with delay spread of up to 250 ns.

To efficiently use the spectrum provided in the 5-GHz range, designers of IEEE 802.11a systems use OFDM techniques. OFDM is a unique form of multicarrier modulation. The basic concept is to transmit high data rate information into several interleaved, parallel bit streams and let each of these bit streams modulate a separate sub carrier. In this way, the channel spectrum is passed into a number of independent, non-selective frequency sub-channels for transmission between wireless NICs and access points.

The OFDM modulation technique is generated through the use of complex signal processing approaches such as fast Fourier transforms (FFTs) and inverse FFTs in the transmitter and receiver sections of the radio. One of the benefits of OFDM is its strength in fighting the adverse effects of multipath propagation with respect to intersymbol interference in a channel. OFDM is also spectrally efficient because the channels are overlapped and contiguous.

OFDM is well-tested and has been adopted by a number of standards bodies for several applications, including a wired global standard for asymmetric digital subscriber line (ADSL) and for digital audio broadcasting (DAB) in the European market.

To complement OFDM, the IEEE 802.11a specification also offers support for a variety of other modulation and coding alternatives. For example, the standard allows engineers to combine BPSK, QPSK, and 16-QAM modulations with convolution encoding (R = 1ý2 and constraint length seven) to generate data rates of 6, 12, and 24 Mbps. All other combinations of encoding rates, including R = 2ý3 and R = 3ý4 combined with 64-QAM, are used to generate rates up to 54 Mbps, which are optional in the standard.

Packet Data Frame

During development of the 802.11a standard, the IEEE 802.11 working group carefully optimized the PHY for traffic transmitting multimedia content such as streaming video. The packet date frame defined in the 802.11a specification consists of the PHY header, PHY convergence protocol (PLCP) and the payload (PSDU). This is similar to the structure used in the IEEE 802.11b specification.

The first field of the PLCP header is called the preamble.

The preamble consists of 12 symbols, which are used to synchronize the receiver. The second field is the signal field. The signal field is used to indicate the rate at which the OFDM symbols of the PSDU payload are transmitted.

The PLCP header is always BPSK modulated and convolution encoded at R = 1ý2. The PSDU packet payload is modulated and transmitted at the rate indicated in the signal field. This rate is variable from 6 up to 54 Mbps. The structure of the packet data frame is illustrated in Figure 3.

On the horizon

On the 5-GHz front, we can expect to see first generation 802.11a products enter the market beginning in 2001 in the form of chipsets, wireless NICs, and access points, with infrastructure expanding in a manner similar to its predecessor IEEE 802.11b. Later in 2001 we can expect to see the Wireless Ethernet Compatibility Alliance (WECA) begin interoperability and compliance testing of 5-GHz 802.11a products.

About the Author

Bob Heile is chairman of the IEEE 802.15 Wireless Personal Area Networking Working Group. He is also director of wireless opportunities at GTE. Heile holds a PhD from John Hopkins University and can be reached at bheile@bbn.com

Thank you to Al Petrick, Vice-Chairman of the IEEE 802.11 committee, for providing the technical details behind the IEEE 802.11a for this month's column.

Figure 1:5-Ghz frequnecy spectrum allocation
Figure 2: IEEE 802.11a frame format for 5 Ghz
Figure 3: IEEE 802.11a channel scheme

Return to the Table of Contents