Consumers and businesses love wireless networking, but the latest version of 802.11 “Wi-Fi” will leave many scratching their heads. With a confusing array of optional features, most of which are disabled by default for compatibility, today’s 802.11n products will not deliver the performance gains promised by vendors and expected by consumers. Let’s take a look at the key components of “Wi-Fi n” and consider what is included and what is not.
A Little Wireless History
Also see Jennifer Huber’s post, The History of Wireless Part One
802.11n is the fifth major revision to the 802.11 wireless Ethernet spec marketed popularly as Wi-Fi. I have been interested and involved since the very start, implementing a dead-end Raylink FHSS 802.11 network in the 1990’s, and skipping 5 GHz 802.11a before deploying 2.4 GHz 802.11b and faster 802.11g networks. Today, I use a mix of 802.11b and new 802.11n devices in both the 2.4 and 5 GHz bands.
Each of these upgrades saw a major performance boost, and each also saw broader adoption by consumers and businesses. Basic 802.11 was a revelation in mobility in the 1990’s but 2 Mb/s couldn’t support real work. After its ratification in 1999, many corporations deployed 802.11a at 5 GHz to avoid the crowded 2.4 GHz band and boost theoretical performance all they way to 54 Mb/s. At the same time, consumers began snapping up standardized 802.11b “Wi-Fi” devices, despite the crowded confines of the 2.4 GHz band and mediocre 11 Mb/s performance.
Introduced mid-decade, 802.11g would become the catalyst for a wireless networking revolution. It matched 802.11a with 54 Mb/s peak performance but was compatible with 802.11b in the 2.4 GHz band. And “wireless g” was widely implemented from consumer and business devices to public hotspots. Home deployment of mixed 802.11b/g access points is common, with an easy majority of broadband Internet subscribers sharing their connection as I do, using a single Wi-Fi router.
But my home office network is nowhere near as advanced as the corporate wireless LANs that are the workplaces of my friends from Wireless Field Day. Most are rapidly adopting 802.11n and deploying its performance-enhancing features. They use “mesh” access points for mobility and deploy wireless controllers to coordinate the network. And they sniff the airwaves for interference and rogue devices.
Four Key Enhancements in 802.11n
802.11n was designed to unify the Wi-Fi world and bring new levels of performance. It includes both 2.4 and 5 GHz radio spectrum as well as multi-stream “MIMO” radio capability, wider channels, and frame aggregation. 802.11n is even backward-compatible with 802.11 a, b, and g networks.
The 802.11n specification adds the following four key areas of improvement:
- Frames and symbols are tuned for better performance. This includes aggregation, which reduces the overhead of data transmission by combining multiple MAC Service Data Units (MSDUs) or MAC Protocol Data Units (MPDUs). Another optional enhancement is a shorter guard interval between symbols.
- Some 802.11g devices included multiple antennas, but true multi-antenna and multi-radio MIMO capability is new in 802.11n. “Wi-Fi n” can combine multiple data streams in the same channel to double, triple or even quadruple throughput using spatial division multiplexing (SDM), but most devices use just one or two radios, blunting the performance benefit. Space-time block coding (STBC) is another option to improve MIMO performance.
- 802.11n is compatible with both the 2.4 GHz band common to consumer Wi-Fi devices and the 5 GHz 802.11a band. When operating at higher frequency, “wireless n” has more channels to choose from and will experience less interference from Bluetooth, portable phones, baby monitors, and other scourges of the 2.4 GHz range. But many 802.11n devices do not operate by default at 5 GHz, and some lack this capability entirely.
- 802.11n can also use wide 40 MHz slices of spectrum, doubling performance but potentially interfering with previous 802.11 devices that used just 20 MHz at a time. This potential for incompatibility has greatly restricted implementation of this feature in the 2.4 GHz band, and such devices are required to support coexistence mechanisms. A new “green-field mode” is optional for high-throughput (HT) networks.
It is difficult to be all things to all people, so the standards bodies made many of the advanced features of 802.11n optional. This means that many “wireless n” products do not support performance-enhanzing features like multi-streaming and 5 GHz operation. In fact, many popular devices offer little more than basic 802.11n compatibility.
Even more troubling, most consumer-grade access points disable the features they do support in the name of compatibility. Buy a brand-new “Wi-Fi n” router and it probably won’t use the 5 GHz band or 40 MHz channels, and is unlikely to contain more than two radios for MIMO and SDM to exploit. Both the device and access point must support each feature to reach maximum performance.
The Current State of 802.11n
Today, many 802.11n access points support both 2.4 and 5 GHz, sometimes with a single radio but often with two. Some also include two or three antennas and can support two or three spatial streams. Properly configured, these access points can reach 300 Mb/s in throughput, but are often limited to 72 Mb/s in their out-of-box state. Many vendors specify a default configuration with 20 MHz channels and 2.4 GHz for compatibility with “wireless g” devices.
Some consumer access points (and most business ones) can be configured for full 802.11n performance in the 5 GHz band while at the same time maintaining an 802.11b/g network at 2.4 GHz. These are known as “concurrent” or “simultaneous”, as opposed to “selectable” dual-band products that only support one or the other frequency range. These offer the best of both worlds, pushing Wireless-N performance without sacrificing “wireless g” clients. But this requires the expense of multiple radios and more-complicated software.
Client devices are flexible and limited at the same time. Cost, space, and power constraints mean many devices do not support 5 GHz frequencies and multiple spatial streams. But those that do will often “fall forward” to improve performance when paired with an appropriate access point.
The latest MacBook Pro, for example, includes a 3×3:3 radio and can hit 450 Mb/s if a capable access point is available. But most PCs and tablets have just one or two radios, and some are restricted to 2.4 GHz as well. Mobile devices like my own iPhone 4 and Eye-Fi camera card are severely limited, with 802.11n included mainly for compatibility rather than performance. The iPhone 4 and Eye-Fi card are 2.4 GHz-only and 1×1:1, the iPad 2 is 2.4 or 5 GHz but still 1×1:1, and the AirPort Express I use is dual-band 2×2:2.
Buyers of 802.11n wireless network equipment should not assume they will see a great benefit right out of the box. Most will have to enable by hand a high-performance configuration including wide channels and 5 GHz operation. And some client devices may never reach the levels of performance expected by consumers due to hardware limitations.
It is disappointing that enhanced performance is disabled by default. But perhaps this will change as 802.11n-compatible client devices become more common.
Also note that using WEP or TKIP encryption disables 802.11n data rates; if you have an SSID using WEP or WPA1, you’re going to top out at 54Mbps. You can only get N data rates if you use WPA2 with AES, or if you disable encryption altogether.