In-Depth
Five Things You Need To Know About 802.11n
Five facts to help you navigate through the new standard.
By Joe Epstein
People often say that wireless networking will someday replace the hard wires that connect people to the network, but for years the technology supporting wireless LANs wasn’t considered sufficiently powerful. Instead, WLANs were perceived as a good supplement to wires, providing mobility and convenience for employees, but with the understanding that users would plug back in to Ethernet as their primary connection. WLANs were simply too slow, too hard to deploy, and too sensitive to the number of users to really be relied on.
With the recently ratified wireless networking standard, dubbed 802.11n, there are new opportunities to change that belief. The latest generation of wireless technology that underlies Wi-Fi was ratified in September, opening up opportunities for wireless networking. Networks based on this new technology can run as much as seven times faster than the previous generation of wireless could -- to over 200Mbps. Along with this greater speed come features that make wireless an attractive option.
If you are building out a new network, or have an existing one, and want to understand what 802.11n means to your environment and users, the five points below highlight the basic, key points that will help you navigate through this new standard.
1. 802.11n can outperform many existing wired networks
802.11n uses a number of new techniques to wireless networking to reach bit rates of 300Mbps -- translating to over 200Mbps of real TCP data after taking overhead into account. Many wireline data networks to the desk are only based on “Fast Ethernet,” or 100BASE-T, providing less than 100Mbps of real data throughput. That gives wireless networks the potential to go twice as fast as wired networks, for the first time ever. Future versions of 802.11n will go even faster, and soon we may see up to 600Mbps for peak bit rates.
Of course, the clients have to share the throughput of one 802.11n access point Some 802.11n network vendors allow the layering of channels in one area, allowing a peak throughput of over ten times in one square foot, exceeding a gigabit of aggregate throughput to split across the users -- more than enough for most networks in the foreseeable future.
2. 802.11n can run your wireline applications at a lower cost
With the higher throughput, it is now possible to move conventional, wired-only applications onto wireless. Medical applications, financial applications, warehouse inventory and shop floor management software, point-of-sale systems, and applications that were too business-critical for legacy, low-throughput WLANs are now being deployed on top of 802.11n.
The idea of a school doing all of its teaching and testing online may not be new, but to have all students connected online in the classroom is, and 802.11n makes it possible.
New applications are also being created just for wireless to take advantage of the blend of capacity and mobility that 802.11n provides. Hospitals are migrating chart and prescription applications to wireless and have started deploying “robo-docs” -- robots with a camera and a screen that can do rounds and visit patients with the actual doctor seated a thousand miles away.
There will still be wired ports for some time to come, but their role will change as wireless begins to share the title of primary network. The fact that wireless networks are less expensive to install -- far fewer cable pulls and ports -- is making sure of that.
3. 802.11n is not like its predecessors, 802.11a/b/g
Compared to previous 802.11a/b/g wireless networks, 802.11n marks a radical departure. It is faster and gets its higher speed in a unique way. This means that there is a learning curve that administrators of wireless networks must go through to understand what 802.11n can do and how it works.
The added speed is possible by using radio waves in a different way. Instead of relying on a straight path from the access point to the client (passing through walls as needed), 802.11n relies on signals that echo off of every surface. This concept, called multipath, means that 802.11n uses a different set of rules to predict when it will perform optimally and when it will have to back down to lower throughputs. For example, tough-to-cover spots in legacy networks can become some of the best spots for performance with 802.11n.
Multipath directly impacts the tools used to plan for, deploy, and manage a WLAN. New tools, based themselves on new techniques, have to be created. Prediction tools, such as RF planning systems and networks that use automatic RF configuration, are the hardest hit, because how well 802.11n will operate cannot be predicted; performance depends on how signals bounce off surfaces. New tools and systems are emerging that can solve the problem a different way, providing predictability by replacing old methods (such as dynamic transmit power control and dynamic channel selection) with robust methods specifically for 802.11n.
4. Device diversity is much higher in wireless than wires, however, so keep your eyes open
The actual technology used in 802.11n is more advanced than that in Ethernet. The greater power each client has provides a greater set of options and places for behaviors to differ. Therefore, the diversity in client capabilities is much greater in wireless than it was for wired connections. Most Ethernet ports can be thought of as essentially the same -- the only important question is whether it is 10Mbps, 100Mbps, or gigabit -- but 802.11n devices support a far greater number of features and can do so quite differently.
The good news is that nearly every client is certified to be interoperable by the Wi-Fi Alliance, an industry organization dedicated to this cause. In fact, they have been certifying for over two years now, first based on an earlier version of the technology but now using the ratified standard. This has ensured that all of the 802.11n devices work together.
Optimal performance still depends on the feature set of each device and how the devices makes use of these features, which, in turn, depends on the particular model of wireless adapter and driver version. For example, in the two years that 802.11n-compatable devices have been commercially available, Intel has sold three different types of Centrino 802.11n adapters: the 4965, the 5100, and the 5300. Each one supports a different peak data rate and a similar, but still substantially differing, set of features. The same holds for other manufacturers.
There is a solution to this, embodied in the last thing you need to know about 802.11n.
5. It is now possible to build a switch-like network with 802.11n
Previously, wireless LANs were a free-for-all, where every device could do what it wanted -- within boundaries -- with the network unable to prevent such activity because legacy wireless was purely hub-based. Wireline Ethernet solved its own free-for-all problem, launching edge networking into the mainstream with the creation of a switch. Techniques have been added to 802.11n that now make it possible to build a switch-like wireless network.
Each wireless device can be segregated into its own wireless “port,” where its ability to affect the quality of the network for all of the other devices is nearly eliminated. Switch-like wireless allows the same methods of bringing the network under control for Ethernet to be used with 802.11n.
Currently, most wireless vendors still sell hub-based 802.11n, but switch-like 802.11n is now available, and it is important for those building wireless networks to seek out switch-like technology, and not just build a faster hub-based network.
The Last Word
Together, switch-like 802.11n, with its higher throughput, application-enabling mobility, and powerful client devices, can finally create a network that can replace Ethernet at the desks and in the halls.
Joe Epstein is the senior director of technology at Meru Networks. You can contact the author at jepstein@merunetworks.com