A Win-Win Situation

What's worse is that, without Microsoft handing me subject matter on a silver platter (namely bugs galore), I may have to get off my butt and do some serious research to find subject matter for future columns. Oh, the inhumanity of it all!

After issuing products that were so unreliable that even the most pro-Microsoft writers bashed them in the press, Microsoft has finally issued enough patches to make the Windows 2000's server versions work well enough that I no longer need to refrain from selling Microsoft's server software on either ethical or self-preservation grounds.

Ethics were involved in that I was unwilling to foist what I consider an incredibly buggy and inherently unreliable server operating system, Windows NT, upon them. Yes, we have worked for some time with NT servers in mixed NetWare/NT environments, but I was never the one to actually sell the NT server to the client. I have seen too many NT servers fail repeatedly even though they were set up by people with extreme expertise in NT. Even the national press constantly publishes that it is considered "standard" to constantly have to reboot NT servers due to bug-induced crashes.

I have not been persuaded by arguments of colleagues that I should push NT Server over NetWare because of the money involved in the additional (presumably expensive) hardware needed, the extra installation and setup costs, and, especially, the never-ending stream of maintenance fees involved in keeping them running.

A secondary consideration (well, maybe really primary) is that I don't want to get blamed, or even sued, when the server proves unreliable. Neither my reputation nor my wallet can afford it. I have already had enough trouble with extreme negative feedback, sometimes to the point of nonpayment, by clients unsatisfied by the reliability of the Windows 95/98 family of products.

In light of this, I have been anxiously waiting (and waiting, and waiting . . . ) for the day when I can offer Microsoft's Windows 2000 as a reasonable alternative to Novell's NetWare. I fully admit that a major part of my desire to offer Windows 2000's server versions is that, first, Novell has repeatedly proven over the years that, while it made, and still makes, the best software available (despite the fact that I nitpick tem regularly), it has not been able to market its products in anything approaching an effective way. It almost seems like they are trying to hide their products from the press, and, in turn, from prospective customers.

Second, Microsoft has proven so effective in marketing their products that the companies that have traditionally been used as examples of great marketers, Procter & Gamble and IBM, could now be considered rank amateurs in comparison to Microsoft. If Firestone was run by Bill Gates, their sales would be soaring right now, and, in light of the various investigations going on, people would be protesting the government's interference in Firestone's business. They probably would have had to hire extra "temporary" employees, and would have had to find creative new ways to underpay them.

Perhaps marriage has mellowed (or at least worn) me out, but, now that Microsoft has caught up with the most critical patches and fixes (emphatically NOT including the now-withdrawn mess known as Service Pack 2), I consider Windows 2000's server versions to be both reliable enough and, as importantly, effective enough, for my company to sell. For me, being willing to sell a product is the ultimate complement.

Reliability isn't the only issue when comparing Windows NT to Windows 2000. In fact, I am somewhat amused by the fact that reports are coming out that say that people experienced with Novell's NDS (Novell Directory System), but not with Windows NT, are having an easier time designing, installing and managing Win2K servers than experienced Win NT folks. This sounds quite similar to the turmoil created among NetWare 3.x experts when NetWare 4, with NDS, was released. To Novell's dismay, there are still thousands of 3.x sites out there run or serviced by people who have still not gotten a handle on the concept of directory services.

Now, Microsoft may be in the same boat. They are using various carrot and stick (mostly stick) methods to convince people to abandon NT, and, especially, to redirect their ongoing educational efforts away from NT and toward Win2K. They are caught in the same chicken versus egg cycle that has hampered Novell's products. Which comes first? Will Win2K be purchased by enough customers to attract both software developers and trained technicians? Or will there be enough software and trained technicians available so as to attract potential Win2K customers?

Now, as to advice on what to do, it depends upon your individual situation. If you have NetWare and are not being forced by either management demands or by application developers pressure, then stick with it. You've already got the best. In fact, if your software developer is insisting you switch just to make them happy, then you might want to reevaluate your application software vendor with an emphasis on finding one with a more customer-friendly attitude.

If you already have an NT server, your only restrictions are the availability of Win2K compatible versions of your server-based software and your budget. If your software developer hasn't yet jumped on the Win2K bandwagon, then a reevaluation might be in order. In this case, you would be looking for one that keeps their products up to date. Basically, the more you are impacted by the unreliability of NT, the more urgent your conversion becomes. If your current LAN vendor can't handle the switch, especially because they can't understand the concept or the benefits of directory services, find one who can. Of course, this advice also applies to current owners of NetWare 3.x and below.

Linux is also a choice, especially for those whose personnel have UNIX experience, but I am not yet anxious to unconditionally jump on that bandwagon. I'm waiting for a shakeout of Linux vendors, so I can see who wins the version and interface battles. I'm also waiting to see better support options for what is currently a self-supported operating system. The support companies that I liked the best have all gone out of business!

An up and coming option is to skip a traditional network server altogether and go with a stand-alone NAS (Network Attached Storage) device. I have been selling standalone NAS devices for a while and am currently gathering information for some future articles on them. Typically they are based on a hidden Linux kernel, but are operated through a browser-based interface so that no Linux knowledge is required. I think that this option may eventually overcome all NOS-based (Network Operating System) choices for many categories of users.

Now I have to get some smelling salts to revive my staff. The shock of reading this pro-Microsoft column was just too much for them.

©2003, Wayne M. Krakau