Abstract

In the world of distributed organizations, business travel, and telecommuting, today's remote users require reliable, efficient access to key corporate applications. The ability to share electronic documents, access customer databases, and run enterprise applications has become a necessity for many remote users.

Attachmate Corp. commissioned The Tolly Group to benchmark the performance of their Remote LAN Node (RLN) version 4.0 product suite, that includes an integrated application server, along with two other remote access products that do not utilize an application server: the Shiva LANRover version 3.5, and the Microsoft Remote Access Server (RAS) supplied with Windows NT Advanced Server version 3.5.

While the RLN Server is used to provide network extension via a remote node connection, the RLN Application Server can provide a potential performance boost for data intensive applications. Additionally, RLN improves upon the limitation of earlier single- PC remote control solutions that require a dedicated PC on the LAN for each remote user, by supporting multiple users on a single PC running a multi-tasking operating system.

Tolly Group Report #5277 (July 1995) compares Attachmate Remote LAN Node (RLN) v4.0 with integrated RLN Application Server v2.3 to Shiva LANRover v3.5 and Microsoft RAS v3.5, timing real end-user tasks over 14.4 Kbps links using cc:Mail, Microsoft Office, Access, SQL Server, and 3270 emulation.

Why RLN is faster. Running the application at headquarters means only screen/keyboard/mouse updates traverse the link, avoiding whole-file transfers; once sessions start, the RLN Server+Application Server combo consistently outperforms the others across tasks.

Email (cc:Mail). Combined attach+send took 73.05 s on RLN vs 155.09 s on LANRover and 186.33 s on RAS; combined open+save was 74.23 s on RLN vs 107.19 s and 118.21 s.

Microsoft Office. With files on the HQ server, RLN opened/saved a 25 KB Excel file in 2.90/2.86 s; LANRover took 37.76/48.46 s, RAS 76.64/93.76 s. PowerPoint open/save was 8.12/5.96 s on RLN vs 25.46/23.46 s (LANRover) and 54.91/48.77 s (RAS). Word open+page-to-end/save was 20.76/2.57 s on RLN, 21.47/68.10 s on LANRover, and 69.13/257.60 s on RAS.

Microsoft Access. Query results of 25–100 KB completed in 14.38–32.79 s with RLN vs 51.82–270.22 s on LANRover and 93.92–235.02 s on RAS.

SQL Server. A 195-row query returned in 3.78 s using RLN Server vs 8.12 s on RAS; LANRover couldn’t run with SQL 6.0 Beta.

Mainframe access. RLN matched LANRover on 3270 responsiveness (2.07 s vs 2.18 s per XEDIT page-down). RAS wasn’t tested for host access due to NetWare SAA limitations.

Bottom line. By centralizing application execution, RLN cuts remote task times dramatically while maintaining parity for host-based 3270 use, making it the strongest choice here for low-bandwidth remote office productivity.