There are circumstances where 802.11n devices cannot operate at their maximum capable data rates. There are various reasons why this occurs. This is the list of factors that affect 802.11n throughput:-
When 802.11n clients operate in a mixed environment with 802.11a or 802.11 b/g clients, 802.11n provides a protection mechanism to interoperate with 802.11a or 802.11 b/g clients. This introduces an overhead and reduces the throughput of 802.11n devices. Maximum throughput is achieved in Greenfield mode where only 802.11n clients exist.
Factors such as Channel width, Guard Interval and Reduced IFS (RIFS) play a major role in the bandwidth. Table 1 and Table 2 show how these factors affect the bandwidth.
Clients ability to send a Block Ack instead of individual frame acknowledgements.
MCS Index configured on the WLC.
Proximity to AP—Clients closer to the AP experience higher data rates. As clients move farther away from the AP, signal strength reduces. As a result, data rate decreases steadily.
RF environment—Amount of noise and interference in the environment. The less the noise and interference, the greater the bandwidth.
Encryption/ Decryption—Encryption in general reduces the throughput due to the overhead involved in the data encryption/decryption process. However, advanced encryption standards, such as AES, can provide better throughput when compared to other encryption standards, such as TKIP and WEP.
Wired Network Infrastructure—Bandwidth of the wired infrastructure determines the speed of the traffic to and from the wired network to the wireless clients.
If using an AP1250, change the AP to H-REAP mode for a 5-10% boost.
If using an AP1140, keep the AP in local mode and enable TCP MSS on the controller. Use the config ap tcp-adjust-mss enable all 1363 command in order to enable it.
Disable RRM scanning to prevent any throughput drops when going off channel. This can yield a 1-3% improvement.
Disable RLDP to ensure the AP does not attempt to connect to rogue devices during testing.
- Use a Wireless Controller 5508 as the data plane is superior to the 4404-series.
Various techniques are employed in 802.11n to provide higher data rates and better coverage. This section details the techniques used.
MIMO: In the existing 802.11 a or 802.11 b/g technologies, transmission and reception of data streams usually happen using only one of the antennas. However, in 802.11n data streams can be transmitted and received over both the antennas. This results in a greater number of bits transmitted and received at a given point of time, effective usage of multipath signals which is usually a problem in indoor coverage. This leads to increased throughput and wider coverage. Table 1 shows the data rates of 802.11n currently supported by Cisco. MCS 0-7 are the data rates achieved using single spatial stream (data bits). MCS 8-15 are the data rates achieved using 2 spatial streams, one over each antenna. Note that the data rates are doubled from 8-15. These data rates (0-15) are described as MCS rates throughout this document.
Note: Further higher data rates are planned for future deployments.
Channel Bonding: The amount of data that can be transmitted also depends on the width of the channel used in data transmission. By bonding or combining two or more channels together, more bandwidth is available for data transmission. In 2.4 and 5 GHz frequency band, each channel is approximately 20 MHz wide. In 802.11n, two adjacent channels, each of 20 MHz are bonded to get a total bandwidth of 40 MHz. This provides increased channel width to transmit more data. Cisco does not support channel bonding in 2.4 GHz frequency (802.11 b/g), because only three non-overlapping channels 1, 6 and 11 are available. However, the channel bonding has more relevance in 5 GHz frequency range where you have as many as 23 adjacent non-overlapping channels currently available. Channel bonding is supported only in 5 GHz, for example 802.11a. Table 2 shows the data rates achieved through channel bonding.
Frame Aggregation with A-MPDU: In 802.11, after transmission of every frame, an idle time called Interframe Spacing (IFS) is observed before transmitting the subsequent frame. In 802.11n, multiple packets of application data are aggregated into a single packet. This is called A-MPDU (Aggregated - MAC Protocol Data Unit). This reduces the number of IFS, which in turn provides more time for data transmission. In addition, clients operating in 802.11n send acknowledgement for block of packets instead of individual packet acknowledgement. This reduces the overhead involved in frame acknowledgements and increases the overall throughput.
Decreased Timers: In 802.11n, few timers have been reduced to decrease the idle time between individual frame transmissions.
- Guard Interval (GI): In 802.11, data is transmitted as individual bits. A certain amount of time interval is observed before the next bit is transmitted. This is called Guard Interval. GI ensures that bit transmissions do not interfere with one another. As long as the echoes fall within this interval, they will not affect the receiver's ability to safely decode the actual data, as data is only interpreted outside the guard interval. By reducing this interval, data bits are transmitted in shorter intervals and provide for increased throughput.
- IFS: IFS is less in 802.11n when compared to 802.11.