Abstract

Gandalf Technologies commissioned The Tolly Group to evaluate the LANLine 5225i Remote Ethernet Bridge, with the main focus on benchmarking data-compression efficiency and latency across a 56Kbit/s WAN link. The project compared the Gandalf platform with other remote Ethernet bridges using the same test bed and data files previously used in The Tolly Group’s Data Communications remote bridge testing, in order to provide a directly comparable view of compression performance under controlled conditions.  

The June 1994 report states that the LANLine 5225i achieved the highest compression ratio and the lowest latency of all the remote Ethernet bridges tested. In the compression chart on page 1, Gandalf delivered a compression ratio of 8.42:1, ahead of Magmalink software at 6.8:1, Newport at 6.71:1, RND at 4.26:1, Magmalink hardware at 4.01:1, 3Com NETBuilder at 3.87:1, Retix at 3.73:1, and Microcom at 3.14:1. Tolly notes that compression was calculated by dividing the total effective WAN throughput by the raw link speed, with the lab increasing traffic until total throughput began to decline in order to find the bridge’s maximum sustainable compression performance.  

Latency was the second major metric in the study. Tolly measured the time required for a single frame to pass through the remote bridges and found that the LANLine 5225i had the lowest latency at 6.22ms. The next-best results were Magmalink hardware at 7.1ms, RND at 8ms, Microcom at 16.7ms, Newport at 21.9ms, Magmalink software at 31.8ms, 3Com at 48.2ms, and Retix at 61.1ms. The latency-versus-throughput chart on page 2 shows that Gandalf also delivered the highest total effective throughput while maintaining that low delay, making it the strongest combined performer in the comparison.  

Tolly took steps to ensure the compression results reflected real processing rather than simple repetition of cached frames. The lab generated 140 unique pre-compressed 512-byte frames to exceed the compression algorithm’s history buffer and verified that the caching buffer was not driving the result. The test bed used two Ethernet LANs connected by a 56Kbit/s link through two LANLine 5225i bridges, with throughput measured by a Wandel & Goltermann DA-30 and an Azure LanPharoh, while latency was measured using dual Ethernet analyzers. Product details on page 3 show the LANLine 5225i as a transparent-bridging remote Ethernet bridge supporting one LAN port, two WAN ports, Ethernet interfaces including AUI and 10Base-T or BNC, WAN links up to T1/E1, and management via Telnet, SNMP, proprietary tools, and console access. Overall, the report presents the Gandalf LANLine 5225i as the leading remote Ethernet bridge in this comparison for both compression and response time.  

Solutions tested:

• Gandalf LANLine 5225i — Remote Ethernet bridge tested for compression efficiency and latency over a 56Kbit/s WAN link.  
• Magmalink Series 3000 Compression Bridge/Router (software) — Software-based remote bridge platform included in the comparative compression benchmark.  
• Newport LAN2LAN/MPR — Remote Ethernet bridge evaluated for comparative compression and latency performance.  
• RAD Network Devices RCB — Remote compression bridge included in the benchmark.  
• Magmalink hardware platform — Hardware-based Magmalink remote bridge included in the comparison.  
• 3Com NETBuilder — Remote Ethernet bridge tested for compression ratio and latency.  
• Retix 4810 LAN Bridge — Competing remote Ethernet bridge included in the benchmark.  
• Microcom MBR/6003 — Remote Ethernet bridge tested in the comparative evaluation.  

This replicated a test run by Tolly and published in McGraw-Hill's Data Communications magazine on May 21, 1994.