Reports & Publications

Microcom Inc. LANExpress Remote LAN Access Multi-user Performance Benchmarks

Sponsor: Microcom, Inc. (Compaq)
Microcom Inc. LANExpress Remote LAN Access Multi-user Performance Benchmarks

Abstract

Microcom commissioned Tolly to evaluate the LANexpress Version 1.0 platform, with the main focus on benchmarking multi-user remote LAN access performance against other leading dial-up LAN access solutions. The project measured both aggregate throughput across multiple simultaneous users and single-session throughput for IP and IPX file transfers, using real traffic between remote clients and LAN-based NetWare and UNIX servers.  


The May 1994 report compares Microcom LANexpress 1.0 with four competing products: Access Builder V3.2, Remote LAN Node V2.0, Shiva LanRover/8E V2.0, and a 3Com remote LAN access product formerly known as the Centrum Remote Server. Tolly states that several hundred tests were run, scaling from one to eight simultaneous users. Both traditional V.32bis modems and V.Fast modem technology were tested, with a specific emphasis on the impact of multi-user loading and per-port fairness as organizations expanded remote-access deployments.  


According to the highlights on page 1, the LANexpress platform sustained aggregate throughput of up to 800Kbit/s and single-session throughput of up to 180Kbit/s. In the eight-port V.Fast IP text-file benchmark shown in Figure 1, Microcom’s parallel-port implementation reached 793,804bit/s aggregate throughput with eight active users. The same figure shows Microcom’s serial-port configuration reaching 467,856bit/s, ahead of Shiva at 173,047bit/s, 3Com at 157,515bit/s, and DCA at 82,869bit/s. Tolly notes that Microcom’s parallel port driver produced a major speed increase over conventional COM1 serial connections, but even the serial Microcom configuration led the other products in the test.  


The report also details single-station and eight-station per-port results. In Figure 2, Microcom delivered 149,063bit/s and a 56-second file-transfer time in the one-port active test, compared with 31,711bit/s and 262 seconds for DCA, 31,494bit/s and 264 seconds for 3Com, and 30,387bit/s and 274 seconds for Shiva. In the eight-port active test, Microcom still led with 99,226bit/s per port and an 84-second transfer time, while Shiva reached 21,631bit/s, 3Com 19,815bit/s, and DCA 10,358bit/s. Additional V.32bis single-station testing on page 4 also showed Microcom leading in both IP and IPX averages across three file types.  


The test bed used simulated asynchronous dial-up lines feeding a dial-up LAN access server connected to a 10Mbit/s Ethernet LAN, with a Sun SPARCstation IPX running SunOS 4.1.2 for IP tests, a 80486DX-based NetWare 3.11 server for IPX tests, and a Network General Expert Sniffer for measurement. The report also notes key LANexpress features including integrated remote node and remote control, remote V.Fast parallel-port modems, support for up to eight V.Fast modem server ports, SNMP management through a Windows application, dynamic dial-in and dial-out, multiprotocol support, and dial-in/dial-out security. Overall, the report presents Microcom LANexpress 1.0 as the top-performing multi-user remote LAN access solution in this evaluation.  


Solutions tested:


  • Microcom LANexpress Version 1.0 — Remote LAN access platform tested for multi-user throughput and single-session performance.  
  • Access Builder V3.2 — Competing remote LAN access solution included in the benchmark.  
  • Remote LAN Node V2.0 — DCA remote LAN access platform evaluated in the comparison.  
  • LanRover/8E V2.0 — Shiva remote LAN access solution tested in the multi-user benchmark.  
  • 3Com remote LAN access product — Competing remote LAN access platform included in the test; the report notes it was formerly known as the Centrum Remote Server.