Of course, Trapeze's Mobility Solution isn't just about RingMaster. Security features include WEP, TKIP, WPA and, with a future software upgrade, AES. Security is designed largely around 802.1x authentication--requiring all users to authenticate with a AAA (authentication, authorization and accounting) server to be granted access to the wireless network. Trapeze also off-loads AAA processing from your back-end RADIUS server, which it claims can relieve potential performance bottlenecks. Unauthenticated users are restricted from admission to the network altogether or are allowed a preconfigured base level of access, such as to the Internet. Users granted access via this method may be promoted by the administrator to higher levels of access based on their MAC addresses. There is no support for VPN access or captive-portal Web authentication.
Trapeze's "Identity-Based Networking" features support subnet mobility services by establishing and maintaining individual users' VLAN associations as they roam, usually routing their traffic back to their home MXs. This solution works well for organizations that implement access-control restrictions based on VLANs. CoS and QoS (Class and Quality of Service) are supported as well.
Our testing of the Mobility System included a few setbacks that kept us in the lab late into the night. Although Trapeze brought more equipment than any other vendor, it chose to run through our testing scenario with a beta version of the RingMaster software to demonstrate rogue-device identification features. However, the new software proved a little buggy, and APs refused to accept updated configurations (we wanted to change the AP channel assignments), forcing an eventual cold boot of all equipment. Trapeze assured us the problem has been corrected.
Furthermore, Trapeze's authentication system is tied to 802.1x, and we experienced problems getting it to work under Windows 2000 with all the NICs on our test bed, a necessary prerequisite to performance testing. We eventually solved most of those client problems--and then learned there was an easier way to connect Windows 2000 clients to the system for testing, using a default access configuration.
Trapeze placed in the middle of the pack in performance, with peak throughput of 5.1 Mbps for 802.11b and 23.1 Mbps for 802.11a. Range tests for 802.11b and 802.11a weren't a bright spot, either: For 11b range, Trapeze finished second to last. Its 11a range proved somewhat better, coming in fourth behind Airespace, Aruba and Symbol. Similar to Aruba, Trapeze had better coverage with its 11a radio than with its 11b radio.