Wi-Fi 802.11n: Still Evolving
This vendor-written tech primer has been edited by Network World to eliminate product promotion, but readers should note it will likely favor the submitter's approach.
With 802.11n ratification a distant memory, news reports regarding this giant leap in WLAN capability have also waned. But while 11n has quietly receded into the background, WLANs have crept out of our data-only world and taken flight as full-fledged network platforms.
Platforms, after all, enable the development of compelling applications, right? So, what's more compelling than being able to stream video between your home office server and your 56" LED TV without wires? Or walking into your company headquarters while on your mobile phone and having your call automatically transition from the cellular network to the WLAN to take advantage of significantly better signal strength? It's all thanks to 802.11n -- complex technology that enables conveniences we will soon come to depend on.
Though ratified, 802.11n is still evolving -- not the specification itself, but the equipment capable of taking advantage of it. For the most part, technology is just now beginning to catch up to the capabilities outlined in 802.11n, with still years before there's parity between available equipment and the maximum capabilities 802.11n can offer. Let's take a look at a few of the key technological advances in 802.11n and see how current technology stacks up.
MIMO, or multiple-input, multiple output, is probably the first thing that comes to mind when people think about what's new in 11n. It is the most visible and the most talked-about of the advanced technologies. MIMO uses complex radio frequency (RF) technology that allows multiple data streams to be transmitted over the same channel using the same bandwidth that is used for only a single data stream in 802.11a/b/g.
Two streams deliver twice the data. Three streams deliver three times the data. This is also why 11n access points (APs) have more antennas than the older a/b/g models. At least one antenna is required per data stream, but keep in mind that not every antenna must be used for data, so the maximum number of data streams is limited by, but not necessarily equal to, the number of antennas on the AP.
This is one area where technology is just catching up with theory. The 802.11n specification allows for up to four data streams. Most currently available equipment takes advantage of only two data streams, but equipment is finally coming to market that uses up to three data streams. APs using four streams are still rare.
To take full advantage of the increased throughput the wireless clients must also be capable of operating at the same number of data streams as the APs, and wireless client adapters are a bit further behind the APs, making three-stream capable wireless adapters still somewhat hard to come by from commercial channels.
Channel bonding does exactly that, it takes two existing 802.11 channels and groups them together to form a single channel, with twice the bandwidth. Two times the bandwidth is essentially equal to two times the throughput, so this is another significant feature in 11n.
In the 5GHz band, the channels that are bonded together are adjacent channels, for example channels 36 and 40 in the U.S. In the 2.4GHz range, where the channels are closer together, the bonded channels are spaced several channels apart, so when channel bonding is used in the 2.4GHz band, a significant portion of the available 2.4GHz spectrum is used to service the one bonded channel.
Channel bonding is already widely used. Though effective in both frequency ranges, its use is sometimes not recommended in the 2.4GHz band because it uses so much of the existing spectrum and can cause interference with neighboring 802.11b/g infrastructures. To be safe, reserve the use of channel bonding to the 5GHz channels.
Block acknowledgements are a tried and true method of wired networking. Block acknowledgements allow more data to be sent before the receiving party must acknowledge receipt of the data. This reduces protocol overhead and effectively increases data throughput. Though the throughput gains are not nearly as large as those achieved with MIMO or channel bonding, the efficiencies achieved with block acknowledgements are certainly worth mentioning. Block acknowledgements are also already widely used in shipping 802.11n-capable hardware.
Aggregation assembles data into more efficient packages for transmission across the network. This efficiency comes in several forms. For very small amounts of data, 11n combines data that would originally have been sent using multiple data packets into a single data packet. For large data sets that can't fit into even the maximum allowed data packet, 11n decomposes the data into multiple packets, but treats the packet stream as if it was a single packet, and then uses the block acknowledgement feature. Though these approaches are quite different, the end result is the same -- reduced protocol overhead, which effectively increases data throughput.
Aggregation is a sophisticated 11n technique, but the user need not worry about it. The use of aggregation, and the specific type required, is determined by the hardware itself and is transparent to the end user. Aggregation has been available in shipping 11n hardware for a while, but support for both aggregation techniques has been inconsistent from manufacturer to manufacturer, though this is changing and most new hardware is now supporting both types of aggregation.
Short guard interval
A guard interval is a period of time that is inserted between data transmissions to prevent overlap between the transmissions. Though necessary, a guard interval is essentially wasted time, which means wasted bandwidth. 802.11n introduces efficiencies, which allow for the reduction of the guard interval, cutting the time in half, hence a "short" guard interval.
This efficiency can lead to up to an 11% improvement in overall performance, but keep in mind that the short guard interval can only be used when a 11n AP is communicating with a 11n client. If the wireless network includes a mix of 11n and non-11n wireless clients, the performance improvements of the short guard interval are reduced since the "long" guard interval will be used with all non-11n clients. Support for this capability is growing rapidly in commercially available hardware.
Beam forming is arguably the most complex new technology introduced with 802.11n. In its simplest form, beam forming allows the transmitting device, whether AP or wireless client, to alter the transmission pattern from its antennas to "direct" the data towards the receiving party. It requires the AP and clients "learn" where each other are before employing any beam forming, and if clients are actively moving about (after all, isn't mobility one of the hallmarks of 802.11?) then beam forming becomes more or less useless.
Given its complexity, it's safe to assume that most 802.11n hardware has yet to take advantage of this part of the 802.11n specification. This is changing, however, and some enterprise-grade systems are beginning to include beam forming as part of the feature set. However, the jury is still out as to how effective, and how widely supported, beam forming will become.
Though fully specified, 802.11n is still an evolving technology. Given that, it will take the theoretical maximum throughput from 54Mbps to 600Mbps, it would be unfair to think that leap would be made overnight. WLAN systems capable of 300Mbps are widely available today, with some supporting 450Mbps beginning to hit the market. It will be a bit longer before we're at the 600Mbps mark, and by then even newer specifications that are currently under development will be driving WLAN capabilities much further!
WildPackets Inc. develops hardware and software solutions that drive network performance, enabling organizations to analyze, troubleshoot, optimize and secure their wired and wireless networks.
Read more about anti-malware in Network World's Anti-malware section.