The following text is copyright 1996 by Network World, permission is hearby given for reproduction, as long as attribution is given and this notice is included.

The end of a dream?

John McQillian's ATM'96 conference a few weeks ago was quite a show and exhibited a not so subtle change from the year before.

The attendees went into last year's ATM Year 3 (the predecessor to ATM'96, John having taken marketing pointers from Bill Gates since then) with the fervor of people going to a revival meeting. The conference was going to be a confirmation of ATM's progress towards its inevitable destiny as the universal network. There was more than a little disquiet in the pews when John said that all was not quite as rosy as some would have you believe. The future looked good but the use of ATM for voice in the telephone carriers was not being planned for at the rate necessary to see wide deployment. The attendees, who had come ready to gather up ammunition to take back home and use on the doubters standing in the way of progress, found more blanks than illuminating flares. ATM's future was still in the future, bright and glorious, but still a ways off.

This year I found the incoming crowd of a different mindset, there were more agnostics than believers. More people were coming to find out than were coming to celebrate. Considering the news from Cisco about the cost of 100 Mb Ethernet switches and others about the potential for near-term availability of Gigabit Ethernet, fact finding was a good mode to be in. Not that there wasn't a lot of good news about ATM and impressive ATM equipment on the show floor. Dozens of talks were given that reported on successful ATM trials, progress in ATM standards, security and ATM, new ATM hardware and managing ATM networks. ATM is clearly a technology in the midst of dynamic development with great potential.

Despite the abundance of positive news, I think that by the end of the conference the majority of the attendees would have answered somewhere between 'could be' to 'yes' to the question John asked Thursday noon--"Has ATM lost the desktop?" This premature death, or, at the very least significantly delayed adolescence, of ATM to the desktop comes not just from the slower than many anticipated development of ATM standards but also from the unexpected strength of the competition. Ethernets are familiar technology at many sites, a 100 Mb version of the same technology, running over the same wires, is a easier sell than an entirely new concept. Now that the cost of 100 Mb Ethernet is plummeting (soon to be about the same as 10 Mb Ethernet) and with the realization that raw speed can solve many of the multi-media support issues ATM was touted as addressing, ATM is becoming harder to justify.

During my 'A Conversation with Paul Mockapetris' session Paul said that he thought that the ATM Forum (and by implication ATM as a whole) had "lost the dream". ATM once was this dream of ubiquity, ATM sea to shining sea and beyond. All would be possible in this dream, from blinding fast data transfer to integrated voice, video and data services. Vast cost savings would be achieved by avoiding the requirement for separate parallel networks and parallel separate network support operations. Some still have this dream, but to me it is wishful thinking. There is no magic technology.

In a way it is sad to see this dream begin to evaporate in the light of day. Purity of mission can be great motivator, but the reality is that technology is always changing. The dream should be the continued evolution of capabilities workable across many environments, thus opening up the dream rather than restricting it to a technology island we will quickly pass by.

disclaimer: Harvard accepts no purity before its time (which can be a long time coming). But the judgment of purity, or lack of it, contained herein is mine.