The new ZoneFlex 7363 from Ruckus Wireless sets an aggressive new point for enterprise-class 802.11n access points. With its intelligent antenna management, nice mix of features, and good TCP and UDP performance in hostile airspace, the ZoneFlex 7363 is an appealing all-around wireless LAN solution for cost-conscious businesses.
I tested the dual-band ZoneFlex 7363, which lists for $599, but Ruckus also announced the $499 ZoneFlex 7343, which operates only in the 2.4GHz band.
The 7300 series eschews the bulbous or funky form factors that characterized previous generations of Ruckus access points (like Ruckus' previous-generation 802.11n AP, the ZoneFlex 7962), in favor of a slimmer design. Measuring in at 7 by 7 by 1 inches and 0.88 pounds, the 7363 has a low, unobtrusive profile and comes with slots in the chassis for a wall or ceiling mount. The unit supports 802.3af Power over Ethernet via the device's lone Gigabit Ethernet port. It also comes with two 10/100 Fast Ethernet ports, and it can be custom-ordered with an optional USB port that supports 3G dongles to provide a wireless backhaul (although this port was disabled on my test unit).
The 7300 series can be deployed as a stand-alone access point or as part of a centrally managed network in conjunction with a Ruckus ZoneDirector. I deployed the 7363 in conjunction with a ZD1006 controller, which quickly identified the device once it connected to my network. However, after I approved the new access point for use with the controller, the ZD1006 immediately lost communication with the AP. Administrators should ensure their controllers are updated at least to firmware Revision 8.2 before attempting to join a 7300 series access point to the network.
Like the 7962, the 7363 supports a total of 16 SSIDs (eight in each band), while allowing different security and QOS (quality of service) policies for each. When using in conjunction with a ZoneDirector, by default the first eight configured wireless networks will be replicated to each radio on the 7363, but wireless administrators can organize SSIDs into WLAN groups to apply different sets of SSIDs to each radio.
Like previous Ruckus APs, the 7363 supports Ruckus' SmartMesh wireless mesh to extend the network to places that can't be Ethernet cabled. The SmartMesh functionality automatically configured whether the 7363 was a root node (backhaul) or mesh node, but I could also manually assign the role as well.
Because the 7363 offers concurrent dual-band support, with separate radio chains for the 2.4GHz and the 5GHz bands, Ruckus advertises 600M bps of theoretical throughput, with each radio operating at 300M bps. This is a little disingenuous: For both radios to operate at the 300M-bps data rate (MCS15), both need to utilize 40MHz channels. While Ruckus does support 40MHz channels in both bands, by default the unit uses 20MHz channels, and many 802.11n clients (including Intel's WiFi Link 5100 AGN and 4965AGN embedded adapters that I tested with) do not support wide channels in the 2.4GHz band. Nor is it advisable to typically use wide channels in the 2.4GHz band, given the dearth of nonoverlapping channels. However, Ruckus is not the only wireless LAN vendor advertising 40MHz channels in the 2.4GHz band, as Meru Networks does the same.
Like its predecessors, the 7363 leverages Ruckus' beamforming technology (called BeamFlex) to optimize a client's connection across an array of embedded antennae on a per-packet basis. This process helps maintain the best possible connection despite environmental influences, thereby avoiding paths with degraded signal quality. Whereas Ruckus' higher-end 7962 offered 19 antenna elements and 4,224 signal paths, the 7363 tones it down to 14 antenna elements and 320 signal paths.
To test Ruckus' ability to maintain the best available quality, I eschewed relatively Greenfield testing in favor of TCP and UDP performance tests performed in the cluttered RF space around eWEEK's downtown San Francisco offices. I also paid absolutely no attention to AP placement, putting the 7363 on a desk in a cubicle, measuring performance to a laptop placed 15 feet away (with glass and metal in between) and just over 100 feet away (adding wood and walls to the mix).
To measure TCP performance I used Ixia's IxChariot 7.1 and its High Performance Throughput script, with one client pair simulating a large file transfer from wired network to wireless. At the close measurement point, performance averaged out to around 118M bps in the 5GHz band and 53M bps in the 2.4GHz band. At the longer distance, I saw about 10M bps for the 2.4GHz band and 8M bps in 5GHz.
To measure UDP traffic, I tried out Ruckus' recently open-sourced ZAP test tool (which can be downloaded from code.google.com/p/zapwireless/), which sends a load of UDP traffic from a wired server to the wireless client, measuring and reporting throughput at intervals. ZAP is concerned with the effects of wireless networking on real-time media, like voice or video, so it reports performance for various percentiles to gauge not only the average and best performance, but the worst as well. I've noted the performance for both the 50th percentile as well as the 99th percentile (the worst 1 percent).
In the 2.4GHz band, ZAP reported an average of around 98/68M bps (50th/99th percentile) at close range, and 17.5/7M bps at distance. For the 5GHz band, I recorded 179/140M bps at close range and 24/11M bps at distance.
Senior Technical Analyst Andrew Garcia can be reached at firstname.lastname@example.org.