In late November, 2001, an industry colleague, knowing that I had an interest in the competitive stance between Linux and Microsoft, sent me a document titled 'Partner Guide: Competing with Linux'. This document was produced by Microsoft in June 2001. It appears to be a training course for Microsoft partners who are obviously facing increasing competition from Linux. I believe that running this document through the analysis wringer can tell us much about Microsoft's current view and depth of understanding of Linux and Open Source software and about their planned method of attacking Linux in the trenches of real-world implementers. As the document is likely to be copyrighted, I am unable to present it in toto here, but, through standard fair-use clauses in copyright law, I am able to clip various segments for demonstration, critique and explication. For the purposes of clarity, I'll highlight in bold any text herein substantially or whole derived from the document we are analysing. The document itself is available online to registered Microsoft 'partners' and you can gain access to it here if you have a partner account.

We'll begin with a review of the course objectives: Microsoft states that:

“this course is designed to arm Microsoft's partners who resell or implement Microsoft solutions with the information they need to promote Microsoft-based server solutions in a Linux environment.”

Now, this is interesting in itself. It has become obvious to Microsoft that there are enough sites which run primarily or solely 'Linux environments' that they warrant specific marketing attention. Furthermore, that these sites would be a substantial enough target market for Microsoft invest time, money and energy in attempts to penetrate them.

What Microsoft's stated aim for its partners is to help them:

1) Address their customers' inquiries about Windows 2000 Server versus Linux

This is obviously confirmation that the message about the competitiveness about Linux is indeed getting through to where it counts.

2) Speak to the better business value of Windows 2000 Server versus Linux

If you can't convince potential users of your wares about the quality and cost-competitiveness of your offerings, use vague and non-descript terminology such as 'better business value' to baffle them.

The course notes then go on to define Linux. In the section entitled “What is Linux”, they claim that it:

“is the most successful open source operating system in the market today and has established credibility as a choice in the enterprise for certain server tasks.”

It's nice to see Microsoft finally acknowledging that Linux has achieved credibility in the enterprise, even if they do qualify it with the 'certain server tasks' moniker. The also state that:

“Linux has moved closer to corporate enterprise computing, as OEMs are shipping substantial numbers of Linux on new server hardware and IBM is offering Linux solutions on its entire line of hardware platforms.”

In the same paragraph, Microsoft admits that Linux is also creeping into the embedded space:

“as embedded OEMs and developers use Linux for its open source code and free to low-cost licensing.”

This is in direct contradiction to most of their statements on the utility of having source code if you are a solutions developer.

What follows is reasonably fair summary of the current state of the Linux business, including notes in IBM's pledged to spend US$1Billion on developing Linux apps; core active areas of deployment (ISP/ASP, education, and small business); a re-iteration that not one of the major Linux pure plays has yet to be profitable; an accurate assessment that the install base estimates of Linux vary depending on research firm or analyst. In all, I could sense neither actual nor hinted inaccuracies in their collection of statements. It seems to be the case that they acknowledge that their target audience will likely see through any attempt at mis-information, which is a somewhat altered tack from their burst of anti-Linux publicity and advertising in recent times.

Their next statement is very interesting:

“As an open source project, Linux is not owned exclusively by any one organization and thus presents unique challenges in responding to it as a competitive threat. Additionally, its license model mandates royalty-free distribution. Together these factors make accurate market sizing information very difficult to come by:…”

They then go on to list the factors which contribute to making actual Linux numbers and penetration a tough estimate to procure, and therefore, make Linux that much slipperier an adversary to counter. These include free downloads, frequency of updates, few copy restrictions, cheap cost, and finally, minimal differentiation between server and client versions of the OS.

I believe that the past few years of 'Linux-by-stealth' penetration within the enterprise is something that is a huge tactical advantage for Linux, and something that has not been substantially touted as such in the past. Let's analyse this for a moment. We are in an industry obsessed with marketshare and mindshare. The more 'territory' or 'penetration' your platform or system has, the stronger and more viable it will be (a corollary to Metcalfe's 'network effect.) If you have a platform (i.e weapon) for which no official numbers (or license keys) need be kept and for which, for the most part in historic times, has entered the organisation by stealth, think of how difficult this platform (stealth weapon ??) must be to procure intelligence on; think of how difficult it is for a competitor to counter using their normal market intelligence methodology.

The course notes then go into a reasonably detailed history of Linux, including references to Minix and USENET posts. It includes statements such as:

“Linus Torvalds is still considered the final decision-maker for Linux kernel code changes.”

Once again, the thrust and subtle intonation of these paragraphs is in direct contradiction to the philosophy espoused by Microsoft's media and marketing minions targeting Linux; these earlier pronouncements against Linux have trumpeted an air of project chaos, direction-less development and standard-practice 'evolutionary death-of-code' Fear Uncertainty and Doubt (FUD). It further indicates that the writer(s) of this document under analysis is indeed curious, interested and possibly sympathetic to Linux's growth, with shades of the written flavour of the Halloween documents. Here's an indication of this:

“The last major release of Linux, 2.4, was released in the spring of 2001 several months late. As Linux has grown in popularity, more and more programmers have taken Linux's source code and adapted it to meet their needs. Linus first wrote Linux for the Intel x86 (386 or higher). It is now available for the DEC Alpha, Sun SPARC, Motorola 68000, MIPS, PowerPC and ARM processors.”

The next section deals with Linux's champions (IBM, Sun) and strength of penetration in certain markets; as a platform which spans multiple hardware lines of any given vendor (e.g IBM); as a platform suited to blades (Sun, Cobalt) and in the embedded space, where we read:

“Linux is also making inroads in the embedded space. Embedded developers and embedded OEMs think that Linux is a flexible, low cost solution.”

An analysis of Open Source Software (OSS) is provided next, taking care to explicitly detail that OSS is not “freeware” or “shareware”, which generally do NOT make source code available. Once again, a sympathetic (to our ears) line is given here too:

“Rather, an OSS license guarantees the availability of the source code.”

Now, while we in the OSS community may see this (free availability and code liberty) as a positive, it's appearance within this document, clearly described with no ameliorative shading, should be cause for concern as to possible modes of attack by Microsoft. Their tactic on this may be to muddy the water between the licensing freedoms and responsibility of free software and those user and ISV-created applications (and possibly even user data!) built atop free platforms. We have seen semblances of this already, with the FUD-machine working overtime to try and scare ISVs and VARs from coding closed-source apps for Linux, for fear of being forced to open the source.

What I suspect to be Microsoft's biggest crowbar leverage-point for attacking Linux is in the long-term viability of Linux companies (which is classic, learn-from-the-masters-at-IBM, FUD) so the following section on Linux business models is of great interest. It begins with a solid overview of the various methods by which one may run a pure-play business around Linux:

“Give the OS software away for free & charge elsewhere –develop unique versions of Linux distribution and strongly promote brand”

Fee on Free (Custom Develop) Model –charge for added value to free product

Service Model –support products and sell customization

Training and Education –provide printed materials and classes

Loss Leader –Give away free software for opportunity to sell proprietary products

Hardware Hardware Systems (VA Research, Intel) Hardware Appliances (Cobalt) Hardware Accessories (Cyclades Research & Development Model

Development portals” (by which I presume they mean sites like Sourceforge.net)

A technical overview of the current stable (2.4) kernel is then provided, with the following segue:

“After being delayed for over a year, disrupting distributor releases and furthering Linux fragmentation, version 2.4 of the Linux kernel was finally shipped in Spring 2001.”

It's interesting to note that they view any discrepancies in the kernel releases proffered by the distributions as a form of fragmentation. While not being a kernel-knowledgeable person, I can't comment on the accuracy of their portrayal of Linux's current abilities, but they do take a few digs, viz:

“…

Poor SMP support for I/O operations. … Dynamic process allocation limit. …”

Linux's strength as a server platform is described in some detail:

“Since its birth, the home field for Linux has been as a server OS running on a general-purpose computing platform. In particular, Linux is viewed as a way of exploiting cost-effective and near-ubiquitous Intel-based hardware while achieving the fast multi-user I/O switching capabilities of UNIX.”

Further information is given on running Linux on more scalable hardware, such as on RS/6000 platforms and S/390's.

Next, non-server functions are looked at:

“Despite advances in a number of markets, Linux as a client OS on desktop systems is by and large still confined to the developer community, where Microsoft estimates that it resides on between 100,000 and 200,000 machines worldwide.”

It would be interesting to see how this figure was arrived at. Considerable research and numerous number projections would need to be undertaken, and therefore a good business case must be have been invoked by the powers at Microsoft that had to instantiate this research. It seems that Microsoft take Linux as a desktop threat seriously. This is borne by their following statement:

“The Achilles' heel of Linux is its lack of desktop applications, and the lack of adoption of those desktop applications, particularly office automation applications, that do exist on Linux.”

The notes then go on to outline what they perceive Linux's biggest (non-server) threat to be: Internet appliances: dedicated in function, possibly embedded, possibly rack-mounted, in kiosks or on custom-built hardware. The following statement should be writ in large type on the marketing brochures of the embedded Linux systems vendors:

“In this setting, Linux holds advantages over the single-vendor commercial OS. All of Linux is exposed as source code, while key pieces of competing operating systems are contained in run-time class libraries and other opaque code.

It is possible to trim away unneeded portions of the general operating system, so that the appliance runs with a lighter footprint, and usually much faster as it is optimized for a specific set of purposes.”

What greater confirmation of business competitive advantage does one need, made even more powerful a statement when coming from a direct competitor?

The next section sees a staple in competitive threat analysis: weaknesses and threats. As for Linux's weaknesses, strong emphasis is given to what Microsoft believe to be the non-viability of the open source business model. Some of the points they raise as weaknesses are obviously laughable, and represent some of the few instances of factual error in technical outline within the document, viz:

“Linux lacks key functionality such as a journalled file system and fail-over clustering”

Some of the points are more wishful thinking and possibly a case of burying heads in sand-buckets:

“Lack of applications. … No good service network. Requires self service.”

The one which I find most mirth-making, as a long-time observer of this industry, is the following line:

“Linux adoption and deployment is limited. No future vision with too many current choices.”

What is comedic here is the execrable statement that there is 'no vision' within the Linux fraternity. Coming from an organisation that didn't even grok the Internet until a decade after its industry competitors did, an organisation which has had so many 180-degree reversals of core technological-path direction (don't forget, Microsoft wrote the core of OS/2 1.x, only to abandon it for Win 3.0) and so many jumps from one vision of systems technology future to another (OLE 2.0, COM, DCOM, DNA, .NET, all in the space of a seven year period) By comparison, I have seen little which has changed in the core vision of Linux and related platforms (BSD etc.) And that vision is to build the best, most stable, highest performance OS platform possible, across a range of architectures and then allow everyone else (companies, ISVs, hackers) to build whatever they wanted atop this platform.

Threats are supposedly treated next, but what we get, in reality, is pump-primed FUD. Consider the following line:

“Linux should be taken seriously as a competitor, but there are threats to its continued success. Encourage your customers to consider these as part of the complete Linux picture:”

What follows is a (wish-) list of half-baked statements with minimal or no corresponding context or analysis to back them up. I'll include the list here:

“The development model is likely to fail as Linux grows into a real OS”

What do they classify as a real OS? Surely, Microsoft has been touting Windows as a 'real OS' for over a decade, and Windows is so bereft of most of the features and functionality of Linux as to make their claim laughable.

“Fragmentation – it is already happening”

Rule of guerrilla marketing #1: If you say things often enough, people will start to believe it. The fragmentation bogeyman needs to be roundly quashed by the Linux community.

“Self-destructive license that may not hold up in massive commercial use.”

I guess someone should tell Microsoft about Oracle, which uses the Linux OS in its Oracle DB server appliance, specifically because it's a real OS and because of the commercially-viable license.

Finally, the refrain I've been hearing most from Microsoft's minions these past 6 months:

“If the Linux pure plays fail, will Linux follow?”

The reasoning for the inclusion of this statement within this training document is obvious; it's the only real target that they can aim FUD at and have any real chance of hitting the mark in the minds of many less-informed IT professionals. The ploy of burying the company which competes with them, by whatever means (viz the DoJ vs. Microsoft court case) has been a very effective one for Microsoft these past 10 years. By projecting this competitive stance, and by having achieved a noted track-record of success in attacking commercial entities in past, it establishes a pattern or expectation of being repeated, not only in killing off competitors, but also their product lines, and will leaving that product or platform's users stranded. As we know, this kind of attack is pointless against Linux per-se, as it itself is not a commercial company which can be commercially attacked, through whatever means. There may or may not be casualties amongst the various companies which proffer Linux systems and services, but that, in itself, will never be a body blow to Linux, and perhaps can be seen in the normal light of competitive attrition and commercial natural selection, i.e., a good thing. We have to send this message out to the various industry watchers who might otherwise see Linux as purely the by-product of a company's efforts, and thus at the mercy of direct commercial attack through Microsoft's normal methods. This understanding, spread widely, would be a huge problem for Microsoft. Without the fear of being platforms being orphaned, many users might actually give Linux a spin! In Microsoft's view, this must be unconscionable and can't be allowed to happen. I've covered more of this topic in an older article titled Shoulders of Giants. A paper on the inevitability of Open Source Dominance

Now the spin fun begins. The next section discusses value propositions, namely, Microsoft's value proposition. The marketing people must have kicked in some of the florid verbiage which was inserted in the notes at this point, as the patois changes considerably:

“In the client engagement paradigm, you must position your solutions to the concerns of your audience. Your approach will be modified based on your knowledge of the company's situation, business model, and the position of your audience in the hierarchy.”

Now, I think I know what this statement means: work out the fear harmonics of your clients and ensure that you know how to play to these harmonics with Microsoft's solutions. As you can imagine, this line of argument is core to Microsoft's (and other vendors') approach to selling solutions and one which we should become malevolently well versed in. One sentence in this section bears highlighting:

“…how your Microsoft-based solutions can be positioned to offset potential competitive advantages…”

This confirms that Linux has many 'competitive advantages' to Microsoft's platforms, and that a substantial and co-ordinated effort (on behalf of Microsoft and it's reseller partners) is necessary to counter these natural advantages. Further, these counters to Linux's advantages are not easily perceived or do not 'flow' naturally, otherwise a detailed course outlining them to un-enlightened resellers would not be necessary to devise and deploy.

In a section straight out of Sales-101, recounting the decision-making process as played out by the various generically defined parties (CEO, CFO, VPs, IT Directors, Staff), Microsoft is exhorting its resellers to recognize their customer's core needs in order to better position Microsoft's solutions over Linux, a fair-enough concept. Several examples of how to eek out these needs are given, including identifying the holy-grail sought by salespersons, 'the competitive value proposition', defined as:

“A value proposition is one crucial benefit or, as happens more often, a set of benefits that occur from using a particular solution. The benefits derived from the solution create a compelling business reason for purchasing and using a solution.”

Linux's value proposition is then highlighted, including points regarding Linux being reliable, inexpensive and open source. The proposed response to these positive values is of utmost interest to the Linux business community, as they indicate the Microsoft marketing thrusts that we in turn have to counter. Having your opponent's battle-plans prior to battle does have its advantages, but only if you study and understand them.

Here is Microsoft's response to the Linux cost advantage:

“Linux is definitely cheap, and even free if you are willing to download it yourself, but you get what you pay for. The cost of software is a small component of deploying and managing an enterprise IT infrastructure.”

Firstly, a few factual sleight-of-hands. It is not necessary to 'download' Linux in order to get it for free. One can buy cheap CDs; re-use existing in-hand Linux CDs across multiple (hundreds or even thousands) of systems; acquire CDs from friends or colleagues. Secondly, we keep hearing Microsoft's mantra of the cost of software being an insignificant component the cost of IT. Surely, if this was the case, then Microsoft, which relies almost entirely on software licence sales for its profit, wouldn't be the biggest pound-for-punch profit making machine on earth. The company I work (Cybersource Pty. Ltd. in Australia) for undertook a study into the costs of acquiring Linux and Microsoft-based solutions for organisations of size from 50 to 250 computer-using staffers. In essence, the license cost difference was upto US$1600 per-seat more for Microsoft's solution. This is more than the cost of the computer hardware equipment, network cabling, system installation costs, and also operational running costs for some part of the first year of operation.

The partner-guide document goes on to (amazingly) claim that:

“Microsoft software is already essentially “free” in large projects-typically less than 3% of the hardware/software cost and less than 1% of the total project cost.”

Their report which provides these figures, would be make for interesting reading. I suspect that there could exist some some combination of esoteric or exotically expensive materiel which, when purchased as part of large mega-project, might ensure that Microsoft's licensing costs only a few percent of the cost, but this would be unlike any project I've been involved in over 20 years in the IT industry.

The document does hit upon two points which count in Microsoft's favour with regards to Total Cost of Ownership (TCO): supportability and integration. In essence, there are many more IT staffers presently with broad ranging Windows-platform skills, which apparently plays to Microsoft's advantage. I say apparently, because my company's experience and my own research in this matter indicates that organisations with largely Linux/Unix platforms on client and server systems, seem to require far-fewer support resources than equivalent Microsoft shops. I hope to release more information on this research at a later date. As for the argument of integration, this is harder to counter; by offering an ever-increasing portion of the enterprise platform and application landscape, Microsoft, as a single source provider, can provider a more tightly-knit package to customers. How this can be responded to by the Linux and free software community is a question with no simple response.

Finally, Microsoft responds to the value proposition that Linux is Open Source Code. They fall back onto the misconception of (interoperable platform) choice as a negative factor in IT rather than positive:

“Customers want to have some influence over the direction of their computing platform. The lack of a single owner and well-known decision making process for open source software makes it difficult for customers to influence or guide the direction of features.”

The irony here is that historically a supply through a single vendor, be that IBM or Microsoft, has almost guaranteed that the vast majority of customers have no influence whatsoever or guide in the software provided by the vendor. The more the single vendor provides of the IT industry jigsaw set, the greater the likelihood that this same vendor will use its position of strength in one segment to muscle in on another, quite often by co-opting or morphing its more widely used technologies as a leveraging tool, negatively impacting the users of those technologies through acquisition of functionality which adds little of use to users, but much of tactical use to the vendor. A prime example of this is Microsoft's Visual Basic product, which has been bastardised in so many different ways since it became the tool of choice for the Windows programming hoi-poloi, that it now resembles a Frankenstein-stitched handiwork language, barely recognisable by its users from a mere few versions back. The cost to its core-constituent users, through Microsoft's insistence on using this development tool to further its ambitions in the component, distributed networking and eXecutable Internet fray, is that there have been multiple instances where whole code-bases written in one version of the language, need to be almost totally re-written when a new version which espouses yet another of the vendor's latest contrived twists or turns in platform technology, comes along. In this way, VB programmers, through their pervasive ubiquity and momentum, have more often been treated as disposable pawns in a chess-battle with Microsoft's foes. Which is in sharp contrast to Microsoft's concept that they, unlike open source, offer a consistent vision of future technologiy directions, which they state as:

“There is no certainty that in the long term Open Source Software will evolve to meet the changing needs of the customer and the marketplace.”

They further dip their metaphorical toes into the irony bucket with:

“The open source development model also depends on its users for testing. Windows NT is tested daily on hundreds of OEM configurations, connected to thousands of peripherals, running thousands of different applications.”

Which is obviously why Windows NT is more stable than Linux! Chortle.

Actually, they also cover this topic next:

“Linux is being used for simple tasks such as file/print and static web page serving. Microsoft customers are using Windows NT Server for demanding, high performance, mission critical applications such as messaging, data warehousing, decision support and e-commerce. Less heavily loaded systems with less complex software suites have high reliability.”

The sheer audacity of this statement is quite breathtaking. Windows NT/2000 are industry- renowned as the least robust of any 'enterprise' platforms and surely far far less robust than Linux. The NT kernel's (NT, 2000) lack of performance under complex load is so well known, that most sites that use it do so in single-application-per-server instances. Thus, there will be a separate mail server, separate web server, separate file and print server etc. This goes some way towards minimising those ubiquitous server crashes which occur so frequently; many sites pre-emptively reboot their Windows 2000 servers on a regular basis, to 'refresh' them. This partly accounts for the number of NT/2000 servers installed. If you could serve five core organisational requirements reliably on one Windows 2000 rather than five, Microsoft's server platform numbers would look far less respectable. Further, to state that Linux is only used in 'simple' tasks is an outright lie. More laughable comments follow:

“Four of Microsoft's key OEMs (IBM, Compaq, HP and Data General) now offer 99.9% uptime guarantees for Windows NT Server, attesting to the high reliability of Windows NT. Research on SP4 shows that a majority of Microsoft's customers see Windows NT Server as reliable or more reliable than either Netware or UNIX.”

Some simple calculations indicate that 99.9% uptime amounts to 8.766 hours of downtime a year. With an average reboot of Windows 2000 servers taking perhaps 5 minutes, this amounts to 105 unscheduled reboots per year, which amounts to over 2 per week, which is obviously a gross simplification of uptime numbers, but there you have it. UPDATE: Thanks to Walter van der Schee and Roland Hedayat for a corection with the uptime math. Now, bare in mind, this is considered (by Microsoft) to be a good uptime figure. I smile wryly when I recall Sun's McNally's comment at Microsoft's statement of 99.9% uptime; he didn't consider 99.9% uptime as a statement of guarantee, but as a threat to system robustness! Delivering only 99.9% uptime on Unix or Linux would be considered woeful and regrettable, not respectable and worth touting in marketing documents.

Their final attempt at debunking a Linux value proposition is with respect to Linux's extensive developer network. They begin:

“Linux often uses the catch phrase “built by users for users” but a more realistic restatement is “built by developers for developers.” The Linux development community is comprised of technical hobbyists and UNIX enthusiasts whose idea of usability is a good text editor with which to modify configuration files.”

While this may have been true in 1996, it's certainly not true in 2002. Part of what they attack is the develop/test/release cycle of open source software, namely from the view of an average business user. By falsely positioning Linux as a platform which its users must cosset with copious quantities of technical care and attention, they are attempting, through FUD, to exclude Linux's adoption by the wider mainstream, i.e. their core market. By telling these potential users that one must download and apply source code patches, recompile, link and deploy, they hope to obfuscate (through deception) the far more common and far far simpler package management (DEB, RPM) technologies which have been the core of Linux's software installation for many years. Not only are these package management technologies present in Linux, they fully obviate the need for general users to glimpse or recompile source code and they are also far more reliable software patching frameworks than Microsoft's own hot-fixes. Consider the number of times that one service-pack or hot-fix has quickly been made available to repair the damage done by its precursor. One could assume that Microsoft is cognisant of sophisticated package management technologies existing on Linux. Microsoft are therefore banking on their prospective clientele not knowing this fact, therefore, perpetuating falsehoods about Linux's complexity is in their best interests. What we must do, as a community, is ensure that the advantages of sane and consistent package management, as provided on open source platforms, are de-mystified as a technology and clearly described to that same prospective clientele.

We also need to formulate a unified response to the much touted gut-level truism of “getting what you pay for”, which is invoked within this document as a touchstone for garnering nods of approval from its readership. If you have any ideas, please forward them to me at the address given above. UPDATE: I've received a fair bit of mail on this issue; there seems to be no clear view or consensus amongst the correspondents as to the best way to respond by way of single line catch-phrase. However, I've been toying with these two, which need to be given as a come-back as soon as someone evokes “You get what you pay for”

• “Linux: like the air you breathe, ubiquitous and free”
• “From Microsoft's perspective, there's a sucker born every minute”

The course-notes then cover a number of case-studies in some detail, none of them proffering any compelling reasons as to why Windows platforms were chosen, or, if in fact they were chosen over a serious Linux competitive threat. In fact, in one of the organisations highlighted (ONE/Northwest a non-profit organisation) the statement is made that Microsoft gave the software away for free. While obviously a magnanimous and noble gesture on the surface, one has to wonder why they did that for ONE/Northwest and not any of the million or more other non-profit organisations and institutions worldwide which use their software. Could it be that Linux is becoming so firmly in-radar view to an increasing number of these institutions, as to seriously threaten Microsoft's position and the only way that Microsoft can stay online in these institutions is to give its software away for free?

One humourous and revealing line from the ONE/Northwest case study is this one:

“Now at ONE/Northwest, PCs stay up and running consistently.”

Now, re-read the sentence carefully; unless you're taking care, you're likely to have read the last word as constantly, which is obviously Microsoft's hope, but that's not what they actually said! A PC which is crashing consistently 3 times a day, is still “running consistently”, but it most certainly is not running 'running constantly' These little word games, placed here as simple Sith mind-tricks to ensnare the weak-minded, are always fun, aren't they?

In a section titled “Addressing Anything But Microsoft (ABM)” the document attempts to stave off some of the growing chorus of complaints about Microsoft and it's pursuit of agenda rather than superior technology and satisfied customers. It does so by exhorting its readership to offer methods of making the prospective clientele feel that they are somehow involved in the great Microsoft machine, through the purchase of MSDN and Technet subscriptions etc.

There is also an attempt to step into Linux's shoes on some fronts, through the identification and highlighting of technologies which Linux has historically had, and which Windows is belatedly acquiring in earnest, just to keep up. Their version of playing catch-up here is also humourous, vis:

“Focus on existing Microsoft applications – Windows 2000 has many mature solutions available for managing systems, including new options that are very favorable to Linux administrators (telnet and scripting capabilities).”

While I know of one or two recalcitrants who continue to use Telnet, their numbers have dwindled in recent years. Microsoft is playing catchup with the Linux of five years ago! Obviously, they see enough of a Linux market, to try and speak its terms and entice system administrators across to Windows. Also on the topic of system management comes this gem, which confirms what most of us have known for many years:

“Indeed, recent research has pointed to the fact that there are some very high staffing ratios for Windows NT/2000 administrators to number of servers managed, sometimes as high as 50:1.”

I've never seen a Windows site with 50 servers without a small army of (overpriced) MCSEs! UPDATE: Numerous correspondants have indicated that I've misread the aforementioned statement on staffing numbers. On the surface, this seems to be correct, but I just can't see how they can claim such a high sys-admin staffing ratio per server. If their own research does indicate this, then if there was a way for us to access this research, it would shed considerable illumination on any analysis of Total Cost of Ownership (TCO) calculations for Microsoft's server platforms.

They include another interesting comment in this section:

” Linux is still relying on multiple efforts in their early stages and while some of these projects hold a lot of promise (such as linuxconf), they're still not at the same level as tools like those provided in products such as Windows 2000…”

Some notes on benchmarks follow:

“Focus on the benchmarks – Point out to your customers that TPC is an independent benchmark council and that Microsoft owns many of the top benchmarks published there. Politely ask where Linux is with regards to transaction benchmarks.”

The answer to this final question would be to point out that while Microsoft 'owns' many of the top benchmarks published within their class by the TPC, Linux actually leads the performance table!

One important point which Microsoft make in this document, and which we in the Unix/Linux community should heed is that Microsoft do make some effort in providing processes and documentation which make it easier to achieve the best results possible (no matter how un-satisfactory these may be) from your Windows platform investment. Not enough 'best-practice', operational and implementation guidelines, documentation and checklists have been formulated for Linux, or, if they have been, aren't widely acknowledged and followed. We do have too much of a 'not-invented-here' mentality, regarding system management, amongst the Linux sys-admin community; this holds Linux back in its push into the enterprise, and needs to be addressed.

Copyright © 2001 Con Zymaris. All rights reserved. Used with permission.

---

About the author: Con Zymaris is CEO of Cybersource Pty. Ltd. a long-standing IT & Internet Professional Services company. Con has been using and programming computers since 1979, and using the Internet since 1989 and is an enthusiastic advocate for open-source software libre. While computers were always a passion which morphed into a career, at the University of Melbourne he actually studied Physics. Con is married and has two (very) active and rumbunctious sons.

---

 
This article was originally published on LinuxDevices.com and has been donated to the open source community by QuinStreet Inc. Please visit LinuxToday.com for up-to-date news and articles about Linux and open source.