802.11n is expected to be ratified later this year and to provide speeds of up to 600 Mbps, which is over ten times faster than the existing standards. 802.11n also has other benefits such as increased range and reliability over legacy standards. What should you know about 802.11n? In this article, I will discuss five of the main technical improvements of 802.11n.
The technical improvements that I will discuss include:
- MIMO Antennas
- Spatial Multiplexing
- Channel Bonding
- Short Guard Interval
- MAC Layer Improvements
Technical improvements are only the first in a two part series on "What every IT professional should know about 802.11n". Part two of this series will discuss the main deployment considerations when implementing an 802.11n wireless network.
Perhaps the most widely publicized enhancement in 802.11n is that of MIMO ("my-moe") antennas, which stands for "Multiple Input, Multiple Output". How does MIMO work?
To answer that question, let's look at how a classic 802.11 wireless transmitter operates:
In this case, the signal is sent out of one antenna (represented by solid orange lines). Due to antenna diversity, the signal is received by two antennas at the other end, but only the best signal is processed and sent up to the MAC layer.
Let's compare that to a MIMO antenna structure:
In this case, the graphic shows three receive antennas. The black, green, and red lines above each represent their own signal sent from a separate transmitter. With MIMO, all three signals are received and processed up the stack. This significantly improves the receiver's "ability to hear" and it represented in the graph above by the orange line.
You may hear different implementations of MIMO such as 2x3 and 3x3. The first number is the number of transmit antennas and the second number is the number of receive antennas. If you hear three numbers, such as "3x3x2", the last number refers to the number of spatial streams.
It is again helpful to take a quick look at a classic 802.11 transmitter.
In this scenario, there is no Spatial Multiplexing and only one data stream is sent from the transmitter to the receiver (represented by the orange blocks on either side of the diagram).
With 802.11n and spatial multiplexing, multiple data streams are transmitted at the same time and on the same channel. They are recombined at the receiver using MIMO signal processing. This is represented in the diagram below with two sets of blocks (an orange colored set representing one spatial stream and a navy blue colored set representing a second spatial stream).
Spatial multiplexing doubles, triples, or quadruples the data rate depending on the number of streams. Remember, you may hear three numbers when referring to 802.11n or MIMO networks - the first is the number of transmit antenna, the second is the number of receive antenna, and the third is the number of spatial streams.
Traditional 802.11 channels are either 20 MHz wide (OFDM) or 22 MHz wide (DSSS). Channel bonding combines two adjacent channels, which effectively doubles the amount of available bandwidth.
Channel bonding works best in the 5GHz frequency band, as there is only space for three traditional, non-overlapping channels in the 2.4GHz frequency band. Therefore, there is only enough space for one bonded channel in that portion of the RF spectrum.
With 5GHz, there are over 20 non-overlapping channels, so you can have several bonded channels operating within close proximity to each other without co-channel interference.
Short Guard Interval
The guard interval is the space between symbols (characters) being transmitted. This is often confused with the space between packets, which is the inter-frame space (IFS). The guard interval is there to eliminate inter-symbol interference, which is referred to as ISI. ISI happens when echos (reflections) from one symbol interfere with another. Adding time between symbol transmissions allows these echos to settle out before the next symbol is transmitted. In normal 802.11 operation, the guard interval is 800 ns.
With 802.11n, shorter guard intervals are possible. The short guard interval time is 400ns, or half of what it used to be. Shorter wait time (guard interval) between symbols increases throughput. However, if it's too short, the amount of ISI will increase, and throughput will decrease. On the other hand, if the guard interval is too long, there is increased overhead due to the additional idle time. Using Short Guard Interval increases the data rate by roughly 10-11%.
In normal 802.11 operation, each directed data and management frame must be acknowledged, as shown in the top row of the graphic. This ACK takes the form of a 14 byte packet. 802.11n provides frame aggregation, as shown in the bottom row.
With frame aggregation, up to 64 MSDUs (MAC Service Data Units - essentially layer 2 frames) can be sent at one time. This "super" frame has one physical layer header, then data frames (each with their own MAC header). Once all the data has been sent, a block acknowledgment is sent.
This is more efficient for several reasons:
- A physical layer header does not have to be transmitted for each data frame.
- The block acknowledgment is much shorter than 64 separate ACKs.
- There are far fewer inter-frame spaces as all the data is aggregated into one burst and all the acknowledgments are bundled together.
So there you have it -- the main technical improvements to 802.11n -- MIMO, Spatial Multiplexing, Channel Bonding, Short Guard Interval, and MAC Layer Improvements.
Keep an eye out for part two in this series, where I will discuss the main considerations for deploying an 802.11n wireless network in the enterprise.