Abstract

Gandalf Technologies commissioned The Tolly Group to evaluate the LANLine 5225i Remote Ethernet Bridge, with the main focus on benchmarking compression-driven throughput across a 128Kbit/s WAN link using the Calgary Corpus and comparing the results with four competing remote Ethernet devices. The project was designed to show how effectively the bridge handled different real-world file types, including compressible image data, mixed traffic, and pre-compressed data that offers little opportunity for further reduction.  

The September 1994 report compares the LANLine 5225i, running software version 1.0.0, against 3Com’s NETBuilder Remote Access version 6.2, Combinet’s Everyware 400 version 2.3.1, Telebit’s Netblazer NP version 2.2, and Cisco’s 2503 Series version 9.14. Tolly notes that three of the five tested devices supported compression, while Telebit’s Netblazer NP and Cisco’s 2503 did not. Testing was conducted between two Ethernet LANs connected by a 128Kbit/s link, with throughput measured using an Azure LanPharoh analyzer while a frame generator created the traffic used for the benchmark.  

The report states that the LANLine 5225i delivered the highest total throughput of all devices tested, regardless of file type. In the chart on page 1, the Gandalf device achieved the strongest results for image files, mixed traffic, and pre-compressed data. Tolly specifically reports throughput of 1,314,784bit/s for image files, equivalent to a 10:1 compression ratio, and 312,392bit/s for mixed traffic, equivalent to a 2.4:1 compression ratio. Among the other compression-capable devices, Combinet achieved 5.47:1 on image traffic and 1.77:1 on mixed traffic, while 3Com achieved 8.32:1 on image traffic and 1.58:1 on mixed traffic. The comparative performance chart on page 2 ranks Gandalf at 100%, with Combinet at about 72%, 3Com near 65%, and the non-compression Telebit and Cisco devices each around 41% of Gandalf’s mixed-data result.  

Tolly also took steps to ensure the results reflected actual compression performance rather than reuse of repeated patterns from a history buffer. The lab verified that the compression cache was not driving the outcome by sending 140 pre-compressed 512-byte frames, which exceeded the typical caching window and produced less than one-to-one compression. The LANLine 5225i itself is described as a transparent remote Ethernet bridge supporting AUI, 10Base-T, or BNC LAN interfaces, switched digital and leased WAN connections up to T1/E1, two WAN ports, filtering for MAC, Protocol ID, and SNAP, and management through Telnet, SNMP, proprietary tools, and console access. Overall, the report presents the Gandalf LANLine 5225i as the top-performing remote Ethernet bridge in this Calgary Corpus throughput comparison.  

Solutions tested:

• Gandalf LANLine 5225i — Remote Ethernet bridge tested for Calgary Corpus throughput across a 128Kbit/s WAN link.  
• 3Com NETBuilder Remote Access v6.2 — Compression-capable remote Ethernet bridge included in the comparative benchmark.  
• Combinet Everyware 400 v2.3.1 — Compression-capable bridge evaluated for comparative throughput.  
• Telebit Netblazer NP v2.2 — Remote Ethernet device included in the benchmark; did not support compression.  
• Cisco 2503 Series v9.14 — Remote Ethernet device tested in the comparison; did not support compression.