The drive to revamp the ARCnet local area network topology by early next year is running into resistance from loyal users, whose battle cry could be, ”If it ain’t broke, don’t fix it.”
The ARCnet Trade Association (ATA), sponsor of the revamping effort, is seeking to boost the speed of ARCnet from 2.5M bits per second to 20M bps, increase the efficiency of network transmissions and expand the maximum size of ARCnet networks.
The improvements are aimed at strengthening ARCnet’s image as a viable alternative to Token-Ring and Ethernet local area network topologies.
However, some industry observers have said that by redrawing the ARCnet formula, developers may be throwing away the features that attracted users to ARCnet in the first place. The network’s ease of installation, reliability and lower cost are its primary strengths against competing technologies, they said.
”[ARCnet] is like a [Volkswagen] Beetle,” said Paul Nelson, a networking consultant for Venture Development Corp., an industry-analysis firm in Natick, Mass. ”To change would be a mistake,” he said.
The process of making ARCnet faster and more efficient will probably make it 50 percent more expensive as well, ATA President Bob Wolfe conceded.
However, he added that ARCnet hardware incorporating the new specification will be backward-compatible with older equipment, which accounts for about 20 percent of the installed LAN market.
Companies currently marketing ARCnet implementations include Standard Microsystems Corp., Datapoint Corp. and Pure Data Ltd.
The ATA is planning three major changes for the upgrade, which will achieve the following:
— more than quadruple the capacity of each ”packet” — the envelope in which data is passed from one network station to another;
— permit each network workstation to send out several packets at once rather than one at a time, as under the current design; and
— boost the size of networks from 255 network stations to at least 1,000 stations.
Many ARCnet users, however, classify the proposed changes as solutions to non-existent problems.
”We never had a problem [with the size of ARCnet LANs]. None of our ARCnet networks need to be larger than 255 [stations],” said Susan Farkas, a vice-president in the New York-based Consumer Service Divsion of Citibank N.A., which has about 3,000 ARCnet network stations.
”We found it extremely reliable, and the 2.5M-bps speed never prevented us from doing anything,” she said.
Several users found the 2.5M-bps ARCnet faster than Ethernet, which transmits at 10M bps. Ethernet networks often appear slower than ARCnet networks, despite their higher speeds, because of less efficient transmission, networking experts said.
”It’s quick. It has predictable performance,” said ARCnet installer Steve Adams, a systems integrator at Sympro Inc., of Emeryville, Calif. Adams also lauded the consistent performance of different vendors’ ARCnet boards. ”Ethernet performance tends to vary from vendor to vendor,” he said.
ARCnet technology is still licensed by its creator, Datapoint Corp., which has allowed only two vendors to manufacture ARCnet chips. The two chip makers, in turn, sell to ARCnet board manufacturers. ARCnet’s closed architecture, some say, ensures a consistency among products from different vendors that Ethernet is unable to achieve.
Still, despite ARCnet’s advantages, it suffers from low recognition in the industry. Enter the ATA, which is seeking to discourage users from switching to other technologies in the face of declining Ethernet costs and a beefed-up Token-Ring, Wolfe said.
”They’re not going to be able to compete with us,” Wolfe said of the other technologies.