Each transmission burst in 802.11 requires a guard interval that makes it so transmissions do not interfere with one another. Normally this is an 800ns timer that allows for time between bursts. This gives a safe time between transmissions. 802.11n outlined the idea of a short guard interval, decreasing the timer to 400ns. This gives less time between transmissions, which could potentially cause collisions between transmissions from the same device. In tests done in a closed environment, there was a definite increase in packet error rate. Without SGI, there was little to no packet error. With SGI, we saw a 30% increase in packet error rate. This is a significant increase in error, meaning that far more collisions are occurring. However, the throughput of a BSS using SGI was significantly better than a BSS that does not. While the error increased as predicted, the error did not affect the much higher rates that SGI predicted. Is this true in an uncontrolled environment?
In tests done in a busy office area with several overlapping BSS, there was an increase in error, but yet again, the throughput of the BSS using SGI was significantly higher than that of a BSS with LGI. While SGI does make the WM more prone to error, for the end user they will be ultimately unaffected. The question is: will this be viable in the large scale, especially as we bound more channels making multiplexing time increase?
BEAMFORMING AND MU-MIMO
The 802.11n standard outlined the idea of beamforming or beamstearing technology. It also outlined the idea of RDP which was designed to set up a shared TXOP between STAs. While little to no one implemented this technology in the 802.11n era, 802.11ac has a much brighter future on this front. many chip makers are building beamforming into the newest ac devices. 802.11ac also expanded on the idea of creating a shared TXOP between devices. With beamforming this is possible, allowing for the STA connected to an AP to transmit simultaneously. This means that the WM will be used more effectively, only needing short pauses in the shared TXOP to deal with the Sounding PPDUs to maintain the correct beaming vector.
Beamforming gives each transmission higher gain, meaning further transmission distance as well as relatively high SNR. This is because the directed beam does not incur interference from other devices within the BSS. In the case of an overlapping BSS, issues with beamforming may arise. With a BSS operating alongside one without beamforming, there is the issue of collisions much like before. Since a higher SNR within a beamformed BSS is present, as interference will not be as much of a concern. This will solve issues with increasing channel bounding because traffic is being sent in directed beams, lessening the possibility of overlapping BSSs.
CLEAR CHANNEL ALGORITHMS
While there may be issues with creating limited space on spectra, APs have the ability to work together to equally distribute the channels that are available. 802.11 APs constantly error checks the SNR of the channel they are operating on. If the BSS is centralized on a channel that contains low SNR, then it announces a channel switch and changes to another channel. It then analyzes this channel, which has higher SNR than the current channel. This means that the spectra will be equally distributed in large area. For example, if there are three APs operating in the same spectra, they will operate on separate channels so there is no overlap. This also will account for bounded channels as well. These algorithms will help ensure the spectrum is shared equally by all parties, making the spectra more reliable and keeping its SNR up.
To read more from Jackson Corson, check out his other blogs: Wireless Reliability, Wireless Reliability - Modulation and Wireless Reliability - Channel Bounding.