The Linux movement is thriving, but its goal should be peaceful coexistence with, not the annihilation of, Microsoft...
Windows, said Bill Weinberg, open source architecture specialist and Linux evangelist for Open Source Development Labs (OSDL).
Weinberg will be evangelizing Linux and open source as a conference speaker at LinuxWorld Conference & Expo San Francisco next week. Before that gig, he discussed the state of Linux server and desktop adoption and its rivalry with Windows in an interview with SearchEnterpriseLinux.com.
On your scorecard, how is the technology rivalry between Linux and Windows going?
Weinberg: I would hesitate to characterize the two deployment trends as a rivalry. Rather, co-existence and even co-deployment is just an attribute of the marketplace and of the growing choices for IT decision makers in the market.
IT managers are no longer faced with a stark choice between 100% proprietary stacks versus 100% open source software solutions. Proprietary and open increasingly refer primarily to development methodologies for key software components, not parallel universes.
On one hand, there exists a wealth of open source middleware, utilities and increasing applications with robust deployments over Windows and Solaris. The most obvious example is AMP [Apache, MySQL and Perl/PHP/Python]. While most commonly deployed as LAMP (Linux plus AMP), there is a growing business for WAMP and SAMP, as evidenced in the growth of companies like SourceLabs in Oregon.
On the other hand, Linux and LAMP today constitute a solid and popular platform for mission-critical deployment of workloads based on mixed open and proprietary software, with examples from Oracle (Unbreakable Linux), Cadence, Adobe, Veritas and dozens of other Tier I and II ISVs [independent software vendors].
Has Linux adoption on the desktop side met your expectations?
Weinberg: The Linux desktop, while still a nascent platform, is growing in adoption and deployments. Foremost has been its embrace for technical workstation applications, followed by transaction workers and basic office functions. Adoption has also been strong on dedicated desktop PC hardware used in thin clients, kiosks and transaction terminals. Double-digit growth constitutes the 'good news'; single-digit share is the flipside of the desktop coin.
What barriers to Linux desktop adoption persist?
Weinberg: The biggest challenges that face desktop adoption come from users' expectation of 100% compatibility, interoperability, even full work-alike operation with the installed Windows base.
The Linux desktop is viable unto itself, but is by no means a drop-in replacement for many Windows desktop workloads. Nor should it be. To replicate the Windows desktop experience would constitute an unending paper chase behind Redmond's road map. Rather, what we are seeing is an organic evolution of the Linux enterprise client to meet a self-determined set of business, technical and regional requirements.
Has Linux adoption on the server side met your expectations?
Weinberg: Server-side adoption is one of the two strongest areas of Linux and open source adoption; the other is embedded systems. Server deployments continue to advance at double-digit rates year over year, both organically and through the efforts of organizations like the OSDL and its member companies.
What's been surprising about Linux server adoption? What's been disappointing?
Weinberg: Surprising to many has been the strong adoption in the last three years of Linux in applications and for workloads previously assumed to be the exclusive domain of proprietary systems. Examples include business-critical applications in banking, finance, insurance, retail, entertainment (film) and mission-critical deployments in communications infrastructure, aerospace and defense.
Disappointing, in general, has been the lack of recognition of these successes; specific disappointments include the gaps that persist in some enterprise workloads and the ROI challenges faced by ISVs in supporting Linux as a core host platform.
How will adoption trends for Linux versus Windows servers play out in the next few years?
Weinberg: Sources like IDC peg Linux server growth in the enterprise running at 35.2% revenues and 31.1% unit growth year over year [IDC May 2005]. This growth trend is strong -- Linux has shown 11 consecutive quarters of double-digit growth in enterprise server deployment.
By contrast, Windows server growth is much more modest, at 12.3% and 10.7%, respectively. So, server-wise, Linux growth is continues to boom. The same IDC study shows Linux share of server deployments (and re-deployments) reaching 28% by 2008.
How will adoption trends for Linux versus Windows desktops play out in the next few years?
Weinberg: On the desktop, Linux is still a niche phenomenon. Less than 3% of all desktops worldwide today run Linux, according to IDC, with growth project to top 5% by 2008.
The good news is that Linux desktop unit growth is thereby in the double-digits, compared to an overall flat market outlook for desktop PCs running Windows.
Focusing on shrink-wrap software, Linux and related software also enjoy impressive growth, with the Linux/OSS [open source software] market slated to reach $14 billion by 2008.
Our surveys show that some CIOs consider Linux to still be an operating system for educational and scientific settings. Why does this perception persist?
Weinberg: The most visible manifestation of most technologies is the points of direct human interaction. With an operating system, this human interface is its use on the desktop. If CIOs only look at the total number of computers in their organizations, then they will likely come away with the perception that Linux is best (or only) suited to technical workstation applications because that is where the open source OS is gaining its first waves of enterprise adoption.
Linux has replaced Unix, and also Windows, as the platform of choice for technical desktop applications in CAD, CAE, EDA, graphics and software development. As important as these workstation roles are, they constitute a minority of desktop applications in the enterprise setting, where they are vastly outnumbered by PC-based productivity applications for office workers performing data entry, financial transactions, corporate correspondences, creating marketing collateral, etc.
If a CIO can look past the sheer numbers of desktops in an organization, then he or she will see that Linux already powers the strategic workloads in the enterprise, on workstations and in the data center, while Windows runs 'the rest.'