Originally Published: Monday, 29 November 1999 Author: Ken D'Ambrosio
Published to: corp_features/General Page: 1/1 - [Printable]

Linux: Swiss Army Knife of Operating Systems

"Generally speaking, most of the hype regarding Linux compares its abilities to run a given task against other operating systems doing the same thing. This is certainly a valid perspective, but it doesn't really take into account one of Linux's biggest plusses: its amazing flexibility."

   Page 1 of 1  

Generally speaking, most of the hype regarding Linux compares its abilities to run a given task against other operating systems doing the same thing. This is certainly a valid perspective, but it doesn't really take into account one of Linux's biggest plusses: its amazing flexibility. For comparison, Windows NT, and especially Windows 2000, have many stratifications -- desktop, server, enterprise, etc. While there are certainly different versions of Linux, tweaked to different tasks, the core of ALL Linux is derived from the same kernel. For specifically this reason, any given feature can be added to any given distribution; this can run the full gamut, including multiprocessor functionality, modifications for high-end web serving, and journaling filesystems.

There is another side, though: in addition to kernel modifications to increase performance, there are also high-level, user-targeted differences. Examples would include nifty abilities such as file sharing for Unix and Windows clients, and even "regular" client-side applications like word processors.

So how does this effect an end-user's purchasing decision? In essence, it makes it much more straight-forward. With Windows NT, or most commercial PC-based Unix variants, the buyer has to make certain decisions prior to acquiring the software. "Will this be for a server? A workstation? How many users will it have to be licensed for? What kind of applications software should I buy with the OS?" All of this greatly restricts the purchaser's control of the situation. For one, budgetary constraints may prevent the ideal software configuration from being ordered. Secondly, once the decision has been made, and the software has arrived, it can be awfully difficult to make retroactive changes to meet a changing environment. Bluntly, the end-user becomes locked-in to their choice, with no easy way to scale up or down.

This is where Linux truly shines. Linux, itself, is almost infinitely mutable, able to be modified by a competent system administrator to meet almost any challenge. On top of that, your purchasing decisions can be made both proactively as well as reactively; we all know that we'd love to make our decisions in advance, but we're also all aware that frequently the world just doesn't work that way. In an Open Source business model, though, you can have your cake, and eat it. Most of the money that will be spent regarding Linux would be on the support side -- and, with multiple vendors offerening multiple support packages, it has suddenly become a buyer's market. This is in stark contrast to Microsoft, Sun, and SCO, where the general rule is "Their way, or the highway." Linux support vendors have to live in a sort of Darwinian environment, where, if they don't offer the kind of support their clients are looking for, the clients can always go elsewhere. Try doing that with Microsoft. Even if you can find a different vendor, you know that all of the questions they can't answer off the top of their head go back to the same place: Redmond. "All roads lead to Bill." This is manifestly untrue with the Open Source model.

So, Linux really is a good software metaphor for the venerable Swiss Army Knife, and on several different fronts: you can purchase a copy from any of a range of different vendors, offering different abilities and tools; you can modify the kernel itself to enhance its functionality; you can add applications and utilities to your heart's content; you can have your choice of support vendor. If you ask any buyer in any purchasing department what one of their biggest concerns is, frequently you'll hear, "single sourced product." It is really quite surprising that it hasn't been thought to apply this to software, as well. With Linux, your choices are myriad, and, by definition, can't be single-sourced. Again, this is in direct contrast to most commercial vendors. While it's unlikely that, say, Microsoft could go out of business, experience has shown that they can yank the rug out from under your IT department by forcing "needed" upgrades, withholding support at whim, and just generally being the proverbial 500 lb. gorilla. About a year and a half ago, Microsoft stopped supporting any clients who were running Novell's NDS due to "security" concerns. The hue and cry that ensued was enough to make them rectify their "mistake," but it's clear that a Microsoft client lives and dies at the whim of Redmond.

Which, again, winds us up at Open Source. Of course, there are other members of the OS community: BSD-variant Unix flavors, and GNU's HURD both come to mind. In terms of immediate commercial viability, however, Linux has the momentum, the developers, and the users.

The conclusion is obvious -- you can have your single-sourced, single-priced, single-company-supported "butter knife," or you can go all the way with the multifaceted, "can-do" Linux.

Comments? Email the author of this piece.





   Page 1 of 1