There are two fundamental ways to specify the "speed" of an 802.11 connection. The first way, which is used in all spec sheets and marketing literature and in every standard Wi-Fi driver and management interface, is to specify a connection rate. When 802.11b is said to offer speeds up to 11 Mbps or 802.11g and 802.11a offering speeds up to 54 Mbps the "speed" being referred to is the connection rate (also called the modulation rate). This "speed" refers to the rate at which the 802.11 transmitter is able to send a constant stream of bits. When considering the useful "speed" of TCP/IP data transmission (as would be relevant to measuring the speed of a web page or email message download or an FTP file transfer) the raw burst rate of bits is only part of the equation.

A TCP/IP data transfer involves 802.11 acknowledgements, TCP acknowledgements and all of the individual packet overhead (which includes source and destination addressing, for example). In addition, there are mandatory gaps between each successive 802.11 packet. Ultimately, too, not all 802.11 packets arrive without corruption from environmental factors.

How 802.11n Provide 300 Mbps Connectivity

There are some fundamental engineering enhancements that are part of 802.11n bit and packet transmission. Some of these enhancements are dependent on the environment and will either be active or not depending on device configuration and environmental characteristics. The most talked-about enhancement, "MIMO" (Multiple Input / Multiple Output), where more than one data stream is transmitted at a time (multiple "spatial streams") is a statistical probability and will vary within any given environment. Here's the breakdown of how 802.11 provides 300 Mbps connectivity (remember, too, that the best-case TCP/IP throughput rate will be roughly 66% of the connection rate):

Modified OFDM

• Sub-carriers increased from 48 to 52
• Maximum connection rate increased from 54 Mbps to 58.5 Mbps

Orthogonal Frequency Division Multiplexing (OFDM) is the technique used for representing a short burst of individual data bits using a specific pattern of electromagnetic signals. The data stream is broken down into "sub-carriers" and multiple sub-carriers are transmitted at closely-spaced adjacent frequencies. 802.11g and 802.11a used 48 sub-carriers. Better hardware circuitry and more sophisticated engineering in 802.11n equipment allowed the number of sub-carriers to be increased to 52.

Forward Error Correction (FEC)

• Sender adds redundant data to allow the receiver to detect and correct errors
• Maximum connection rate increased from 58.5 Mbps to 65 Mbps

The structure of transmitted bits is mathematically modified to help the receiver make sense out of bit streams which may have been corrupted by noise or interference. To do this the transmitter adds additional bits in a pre-defined way. The result is referred to as a "coding scheme." 802.11g and 802.11a add one extra bit for every three bits transmitted. This referred to as a 3/4 coding scheme. 802.11n introduces technology that allows a 5/6 coding scheme where one extra bit is needed for each 5 data bits. The enhanced error correction scheme reduces the number of redundant bits needed to compensate for noise and interference.

Shorter Guard Interval (GI)

• OFDM inter-symbol guard interval minimum reduced from 800ns to 400ns
• Maximum connection rate increased from 65 Mbps to 72.2 Mbps

Transmitted electromagnetic signals are reflected off various surfaces as they travel through space to a receiver's antenna. This means that some signal will take a longer path, and hence a longer amount of time, to reach the receiver's antenna. After a transmitter has sent an OFDM burst of bits it must wait long enough to allow the majority of reflected signals to reach any potential receiver within the intended range. In essence, it has to wait for the air to get quiet again before sending another OFDM symbol. In 802.11g and 802.11a this required 800ns but the improved engineering and circuitry in an 802.11n radio allows proper operation after only 400ns.

It's very important to note that use of the shorter guard interval is dependent on the environment. In some spaces a large number of signal reflections may arrive over a long period of time (i.e. greater than 400ns). In this case the 802.11n chipset automatically backs off to the 800ns GI even if the radio has been configured to use the 400ns GI. In almost all real-world environments it will be found that use of the 400ns GI is precluded. Most real-world situations just don't allow for use of the shorter guard interval. Consequently, the data rate can't be increased from 65 Mbps to 72.2 Mbps and this single enhancement is lost.

Channel Bonding (Use of 40 MHz, "Double-Wide" Channels)

• Channel bandwidth is increased from 20 MHz to 40 MHz
• Maximum connection rate is increased from 72.2 Mbps to 150 Mbps
• If the 400ns GI is unavailable then maximum connection rate increased from 65 Mbps to 130 Mbps

802.11g and 802.11a transmit data in a 20 MHz wide channel. At the upper and lower boundaries of a channel's frequency span the transmitted signal's power must drop off to avoid adjacent-channel interference. When two channels are "bonded" to create an 802.11n 40 MHz channel the result is that more than twice the original available bandwidth is created. This is because the space between the original two channels, where the signal originally had to drop off, is eliminated so you not only get double the original bandwidth but you don't loose the adjacent-channel gap that was between the two original channels. That's why the connection rate more than doubles when using 40 MHz channels.

It's very important to note that 40 MHz channels can't always be used. There are severe limitations to the use of 40 Mhz channels in the 2.4 GHz spectrum used by 802.11b and 802.11g. There can be some limitations in the 5 GHz spectrum used by 802.11a. A complete, detailed discussion of 802.11n channel configuration is available here:

READ THE DETAILED EXPLANATION OF 802.11N CHANNEL CONFIGURATION

Spatial Multiplexing (MIMO)

• 802.11n Draft 2.0 supports two spatial streams
• The final 802.11n standard (September 2009) supports up to four spatial streams
  • Each additional spatial stream increases the maximum connection rate by 150 Mbps
  • Two spatial streams creates a connection rate of 300 Mbps
• If the 400ns GI is unavailable then two spatial streams creates a connection rate of 260 Mbps
• A 600 Mbps connection rate is achieved with four spatial streams

A stream of bits is converted into electromagnetic signals. An 802.11b/g or 802.11a transmitter sends these electromagnetic signals out a single antenna. If two or more antennas are present then one is selected at any given instant as the "best choice" for transmission. This is referred to as a "single input" transmission - one antenna sends signals into the air. On the receiving end, a single antenna is used to receive the transmitted bit stream. If two or more antennas are present then they're sampled and compared and the one that is receiving the higher quality signal is selected for use at that instant. This is referred to as "single output" since only one antenna takes signal out of the air at any given moment. "Antenna diversity" is the term for the technique used with most 802.11b/g and 802.11a radios where the receiver has two antennas and selects the best one at each successive moment. Nonetheless, only one antenna is used for reception and the technique remains "single output."

802.11n has sophisticated hardware circuitry and implements software that uses advanced mathematical algorithms to allow a transmitters bit stream to be split and simultaneously transmitted using two (or more) antennas at the same time. Both antennas transmit at exactly the same frequency and on the same channel but the mathematics allows the receiver to differentiate between the two transmitted streams.

Each transmitted stream is called a "spatial stream" because they're differentiated based on which antenna transmitted it and the two antennas are separated in space by a few inches - they're spatially separated. Consequently, multiple spatial streams are being input into the air ("multiple input") and they're then being taken out of the air ("multiple output") by the receiver. This the definition of "Multiple Input / Multiple Output" or "MIMO" (pronounced "my-moe").

For two spatial streams to be separated by the receiver demands that the arrival time of their respective reflected signals be sufficiently unique to allow the mathematical algorithms to work their magic. There's no guarantee that any particular environment will have a reflective nature that is suitable. Moreover, there's no guarantee that a client device will be positioned in a spot where signal reflections will allow differentiation. In fact, sometimes moving a client device only one or two inches can take it from a place where signal reflections do allow differentiation between spatial streams and one where multiple spatial streams can't be separated.

MIMO is a dynamic capability based on the statistical probability that a suitably reflective environment will exist in the very small place where the receiver's antennas are located. In a particular room some users may be able to connect with multiple spatial streams while others may only be able to use a single spatial stream. The situation is totally dynamic, changes from moment to moment, and can't be absolutely predicted.

An optimal 802.11n design attempts to maximize the probability that multiple spatial streams can be used by paying attention to the signal scattering effects in the environment. With a correct 802.11n design it's possible for almost everyone's device in a room to capture multiple spatial streams and realize the associated doubling (or more) of their throughput.

READ THE DETAILED EXPLANATION OF 802.11N SPATIAL MULTIPLEXING (MIMO)