The N Crowd: Wireless N Put To The Test

802.11n wireless networking

The IEEE has yet to ratify the 802.11n WLAN standard, which promises high speeds and increased range over its predecessors. Ratification isn't expected until 2009, but in the meantime all of the major wireless vendors have launched products based on the draft version of the standard over the past year. For VARs, this is new technology that offers tremendous promise in terms of both performance and profit opportunity, but it's not without its challenges.

Aside from the draft status of the technology, solution providers are dealing with new expertise requirements and a lack of mature planning tools.

"There are always nuances with any new technology, but 80 percent to 90 percent of it operates the same as the traditional protocols. There is some new technology embedded, such as MIMO, higher throughput, that people have to learn out in the field," said Brett Rushton, vice president of network strategy and infrastructure at Calence LLC. Calence, a Tempe, Ariz.-based solution provider, has 10 to 20 clients deploying 802.11n-based products.

The lack of mature 802.11n design tools is one of the biggest challenges solution providers face, leaving them to rely on their own field experience to ensure top-level network performance. There can also be snags with power consumption, as the power specification for 802.11n exceeds the original specification for Power over Ethernet, Rushton said.

Austin Smith, CEO of Digital Son, I.T. Services, a Lawrenceville, Ga.-based solution provider, said some customers are still hesitant to invest in 802.11n. But for those that are leery of the technology, an on-site demonstration is usually all it takes to kick-start a sale.

"If they have a need for wireless-N, just demo it. For one music entertainment customer, I brought some gear over and set up a demo. Just dragged-and-dropped a raw music file across the network," Smith said. "It took 30 seconds vs. 35 minutes to an hour-and-a-half. They can see the raw difference in the numbers."

Despite 802.11n's challenges, the channel knows a good thing when it sees it. "There's going to be strong momentum in the channel as wireless moves from being nice to have, to being critical. We're seeing strong interest and strong demand for it," said Nadeem Ahmad, director of global technology for solution provider Dimension Data, Reston, Va. Ahmad is recommending that Dimension Data clients deploy 802.11n as a pilot in preparation. "The final is coming; you don't want to wait," he said.

Next: Methodology Methodology
Last year, the CRN Test Center examined a group of SOHO access points from a variety of vendors based on the draft version of the forthcoming 802.11n high-speed wireless standard (see CRNtech, Jan. 29, 2007). In the meantime, all of the major enterprise players have now jumped on the bandwagon, so reviewers set out to take a look at some of these higher-end offerings.

Reviewers extended invitations to Aruba Networks, Bluesocket, Cisco Systems, Meru Networks, Ruckus Wireless and Trapeze Networks. Only Meru and Ruckus voluntarily participated.

Meru and Ruckus, both of Sunnyvale, Calif., provided products and sent engineers to the CRN Test Center lab to help testers configure the WLAN access points and controllers. Reviewers also examined an 802.11n access point from San Jose, Calif.-based Cisco that was received from an alternate source. No Cisco engineer was present. See the article, Where's Cisco With Wireless N? for Cisco results.

Meru's offering included its AP320 access point and its MC3000 controller. The devices were running Meru software version 3.4.SR3-113. Meru's access point connected both of its radios—2.4 GHz and 5 GHz—but only the 2.4-GHz radio was left on for the main tests. The antennas were configured on a X-Y-Z axis.

The Ruckus setup included its ZoneFlex 7942 access point and ZoneDirector 1000 controller. The Ruckus software version was 3.02. The ZoneDirector is not necessary to configure the 7942 access point but with the help of a Ruckus engineer, reviewers used the ZoneDirector to tweak various settings to optimize performance.

The Test Center selected three notebooks: a Toshiba Portege M700-S7002 with an Intel chipset 4965AGN, an Acer TravelMate 6291 with an Intel chipset 3945ABG and an IBM ThinkPad T61 with a Linksys Wireless N WPC4400N adapter card. The ThinkPad and the TravelMate ran Microsoft Windows XP SP2 while the Toshiba ran Vista. To eliminate unforeseen network problems, reviewers disabled the LAN adapter and only turned on the IPv4 connection on the wireless adapter. Reviewers downloaded the latest Linksys driver and added a Microsoft hotfix for the XP clients to get WPA2-PSK encryption working. Reviewers decided not to install the latest wireless client drivers because they wanted to test under real-world conditions.

Reviewers ran the tests with single access point configurations on Channel 6 and kept channelization at 20 MHz. In addition, testers asked Meru and Ruckus to provide a 2x2 MIMO antenna setup running at 2.4 GHz.

The Test Center evaluated the products on network performance, management features and functionality and channel programs, using Ixia's IxChariot testing software after finding considerably poorer results with SolarWinds' Engineer's toolset.

With IxChariot, reviewers ran downlink, uplink and uplink/downlink tests, each with four sets of TCP socket pairs of traffic. IxChariot was configured to run in batch mode for one minute. Reviewers used the same configuration file for all tests. They tested orientation at 45-degree and 90-degree angles with respect to an access point, and also tested the notebooks with WPA2-PSK security features turned on.

Next: Ruckus ZoneFlex 7942 Ruckus ZoneFlex 7942
The Ruckus ZoneFlex 7942 impressed reviewers with its strong feature set and performance, earning a first-place finish. Though Ruckus and Meru don't directly compete in the market, Ruckus outscored its opponent in areas such as quality and reliability, automation features, ease of deployment, ease of use and price/performance.

Ruckus offers a Web-based management interface to its ZoneDirector WLAN switch, which was used effortlessly to set up the test wireless environment. From there, IP addressing, SSID and security methods are configured. Ruckus' system supports up to four SSIDs (the next release will support up to eight). Total time for initial configuration was about five minutes.

After the initial configuration, the interface opens up into a dashboard view. From the dashboard, detailed information of each currently managed access point is displayed; MAC address, model, IP address, channel and number of clients connected are all listed for each access point. From the dashboard, access points can be remotely restarted.

The dashboard contains a log of recent user activity with a corresponding severity level. Activities tracked include users connected to individual access points, signal strength, user idle time and disconnects. System activity is also logged and detailed info on occurrences with access points are listed.

For a quick view, summarized information is listed on the left side of the dashboard. System overview of the ZoneDirector controller is displayed as are a devices overview and a usage summary.

The interface also has a Monitor view, outside of the dashboard. With this view, a floor plan can be uploaded; access point locations can be graphically represented throughout a facility. This view also gives total number of access points, rogue devices and overall number of clients connected. Generated PSKs and Guest Access passes can also be displayed here.

From the interface's Configure view, access controls can be set for users. Devices can be outright blocked and this is also the area where guest access is granted—access points, users, alarm settings and authentication servers are configured in this view. Ruckus supports a variety of authentication methods; not only is RADIUS supported, but authentication can also be set up against Active Directory and other databases, allowing for the integration of security from a wired network. The company also has a patented dynamic PSK security solution: Each user is assigned a dynamic PSK key that is revocable.

A minor complaint: The refresh rate of the dashboard is automatic, and seemed a bit too frequent. During testing, the constant refresh of the interface caused the browser to stop responding. Reviewers did not see a readily apparent way to control the refresh.

Ruckus' interface is detailed, allowing for complete remote management while at the same time avoiding a cluttered and daunting look. Ruckus' overall performance test results were admirable. Keep in mind that the Ruckus access point technology uses patent-pending horizontal and vertical polarized antennas that are able to coordinate multiple paths as signal strengths change on a client. The antennas are designed for high performance in semi-penetrable walls because it uses beam steering to refocus access to connected clients. Other vendors use omni-directional antennas, where client position makes for a more significant difference in performance.

Ruckus' downlink test results against two 802.11n clients were on average 56.7 Mbps. With the addition of an 802.11g Acer client, the Ruckus access point ran slower at an average of 48.9 Mbps. On the uplink test, the 7942 access point ran faster. The access point scored an average of 65.9 Mbps.

The uplink/downlink test was the toughest because of the large set of data passed between the server and each client. The 7942 access point performed well due to its innovative technology and received an average of 56.6 Mbps, using wireless-N clients only. With the 802.11g client, the 7942 access point received an average of 48.9 Mbps for downlink-only traffic and a score of 66.41 Mbps. On the uplink/downlink, the access point scored 48.9 Mbps.

After changing orientation at 45-degree and 90-degree angles, the Ruckus access point performed equally well. For instance, at 45 degrees with all three clients running, the access point scored an average of 61.1 Mbps. When encryption was used on all three clients, testers noticed a 1 percent to 2 percent drop in overall average performance.

Next: Meru AP320 Meru AP320
Meru's 802.11n offering bested its rival in management capabilities and logging and reporting features, but landed in second place. Reviewers felt the systems were on par with each other when it came to strength of features and profit potential for partners, but Meru trailed a bit in performance and couldn't match the ease of use and deployment of Ruckus.

Through a Web-based management interface, all standard configuration and management options are present. Reviewers found that configuration of the controller and stand-alone access point was quick and easy with the management utility. Within the utility, a variety of authentication options can be enabled, including 802.1X with EAP-TLS, Tunneled TLS, Protected EAP and open authentication. Encryption options include static and dynamic WEP keys (40-, 64- and 128-bit), TKIP and IPSec.

Multiple ESSIDs and VLANs can be generated via the management tool. Security policy management is centralized. Administrators can assign security policies to individual ESSIDs and VLANs.

The interface also allows for the management of virtual cells. Virtual cells are the foundation behind channel layering. Meru has devised a method in which differing channel blankets are created. Channel layering is for better signal throughput and less signal interference.

One of the more unique features of the interface is the management of quality of service and filtering rules. Both standard and dynamic rules can be created, based on applications, stations, uses and port numbers and are applied to clients. Filtering is done via access control by client or access point MAC address. There are also configurable firewall settings.

The Meru's dashboard interface allows for sophisticated and complex management that can seem intimidating to those not well-versed in wireless technology. However, it offers features and configuration options for a true enterprise-level solution.

Meru's solution performed marginally better than Ruckus on some of the tests but trailed on others. Because the clients were in close proximity to the access point, power to its antennas needed adjustment. The access point also required some tweaking. The Meru technology performed on par with the Ruckus access point on the downlink tests.

With only wireless-N clients running, Meru rated an average of 76.7 Mbps for downlink traffic. But when mixed with the Acer 802.11g client, the uplink traffic dropped substantially to 37.3 Mbps. On the ThinkPad in particular, performance dropped significantly.

Meru's AP320 performed well on the uplink/downlink with all the notebooks running. With the Toshiba and the Acer systems, the AP320 scored an average of 44.7 Mbps on the uplink/downlink test.

The uplink/downlink test at 45 degrees with all three clients scored an average of 36.9 Mbps. At 90 degrees, Chariot scored 44.2 Mbps. Without the IBM-Linksys client, the numbers did not improve. There were no noticeable changes in throughput with WPA2-PSK encryption turned on.

Next: The Bottom Line The Bottom Line
After a review of testing results and work with the management interfaces of both products, the scope of vendor offerings became clear: For smaller outfits, Ruckus provides an innovative, vigorous solution that allows for comprehensive, centralized management, while Meru's product is targeted more for midmarket to enterprise-level customers. Management is detailed and feature-rich, but there were still issues with products approaching the full bandwidth 802.11n promises.

Ruckus and Meru have a strong standing in the 802.11n space, but wireless-N is a new field, and vendors are still testing and tweaking their goods. VARs should keep watching this competitive area, as vendors compete, and demand for faster and more scalable wireless access continues to grow.

-Jennifer Hagendorf Follett and Andrew R. Hickey contributed to this article.