Do Not Go Gently Into That Microsoft Night

By Dave Molta  Many years ago, I was certain that my passion for sports, combined with a fair grounding in statistics, would make me a winner at a Las Vegas sports book. But a run of "bad luck" convinced me that there were just too many variables to turn this game into a science. Now, I follow the betting line so that I can root for the underdog in games that don't involve one of my favorite teams. I often wonder how this empathy for the underdog effects my professional judgment as a network and systems manager. We're all called on to make bets on technology, often with information no better (and sometimes much worse), than what's available to sports gamblers. Many network professionals stake out their position early and stick with it, oblivious to reality as their favorites fall by the wayside--long live Token-Ring, Banyan VINES and the Macintosh! Others ride the fence to the bitter end, refusing to take a stand at all, punting strategic issues that beg for decisions. Most claim their turf somewhere in the middle. I can recall my ea rly years in the trenches, when I was an emotional advocate for certain technologies and vendors. I believed that U.S. Robotics had a sure thing with HST modem technology. OK, so that was one that didn't exactly go my way, but my bets on Ethernet, TCP/IP and Novell in the 1980s made up for that. Over time, I've changed my thinking about technology advocacy, partly as a result of experience and partly as a result of an increased sense of conservatism and risk aversion that seems to come with age. You Won't Get Fired For... Many years ago, prevailing wisdom suggested that if you had doubts, go with Big Blue and you won't get fired. But I never felt right following the herd. I bought an IBM-compatible PC for home and installed 10BASE-T before the standard was fully ratified, despite an impassioned case stated by IBM's network expert that the laws of physics would condemn users to poor performance because Ethernet would never run at speeds greater than 10 Mbps. I ignored him. At that time, I was fo cused on two issues: quality of technology and price/performance. There was less of a concern with risk and strategic vendor relationships. Today, all that's changed. I am at once more jaded in my evaluation of new technologies and more careful in my vendor selection. I suspect the same is true for many of you. Some folks would argue that I don't take a clear enough stand on technology directions. But I believe this approach is warranted by the realities of today's marketplace. I offer two examples. A few years ago, we set out to replace our aging campus backbone, which was running on second-generation Cisco routers. The system was reliable, but it didn't have the packet-processing horsepower needed to support higher-speed LAN technologies. The emerging wisdom then was when it came to campus backbones, ATM was the only logical choice. It had high speed. It had quality of service. It was the future. But my network engineering staff saw it differently. ATM was extremely expensive, it was very complex a nd it lacked a proven track record of interoperability. From our viewpoint, it was inconceivable that we would be able to replace Ethernet at the desktop with ATM, so the quality-of-service story wasn't quite so compelling. As far as speed was concerned, 100BASE-T offered far superior price/performance and it was CSMA/CD, something we understood quite well. And we saw this new technology called Gigabit Ethernet on the horizon, and that looked pretty darn good. So, we let the ATM bandwagon pass us by and stuck to a routed backbone supplemented by high-speed Ethernet switches. More recently, we've been fighting the directory services battle. Given our historic reliance on NetWare for file and print services, we decided to deploy NDS as a means of simplifying administration of 15 central servers while providing a higher level of integration with departmental systems, most of which also run NetWare. But our NDS deployment hasn't crossed the invisible line that separates tactical from strategic. As much as I li ke Novell's technology, the strategist in me worries that the hoards who have jumped off the Novell bandwagon may just have the critical mass necessary to condemn NDS to being a niche technology. The general lack of NDS enthusiasm coming from Unix vendors like IBM and Sun or application vendors like PeopleSoft and SAP, combined with Cisco's decision to sign on for integration of its products with Microsoft's Active Directory, makes betting on NDS a little like betting on a four-touchdown underdog to win the game outright. Sure, it could happen, but it's not the kind of wager you jump at without some fairly good inside information.

Other Columnists Corporate View By Robert Moskowitz On The Edge By Art Wittmann