Markets, A. Butler Research Note 17 September 2003 ISV Support Is Key When Choosing a Server Operating System It's the support independent software vendors offer that determines the life span of a server operating system. Yet ISVs don't like saying which they prefer. Gartner reveals their favorites, so you can choose safely. Core Topic Hardware Platforms: Server Platforms Key Issue How will centralized and distributed servers evolve during the next five years? Independent software vendors (ISVs) must be enthusiastic about a server platform if it is to succeed, especially for mainstream corporate server platforms. But ISVs are usually reluctant to tell customers which platforms they are most committed to. ISV enthusiasm will vary over time, as the popularity of operating systems waxes and wanes. Gartner's ISV support profile (see Figure 1) charts the relative viability of server operating systems on two axes: Overall enthusiasm for each operating system. This is a measure of its popularity and market momentum. High enthusiasm typically equates with quick bug fixes, quality performance tuning, knowledgeable technical support and regular new product releases. The relative size of the software portfolio for each operating system. Larger software portfolios are enjoyed by top-tier operating systems, such as Windows or Solaris, and by popular legacy operating systems that mainly sell within their installed base. OS/400, running on IBM's iseries (formerly AS/400) is a good example. Gartner Reproduction of this publication in any form without prior written permission is forbidden. The information contained herein has been obtained from sources believed to be reliable. Gartner disclaims all warranties as to the accuracy, completeness or adequacy of such information. Gartner shall have no liability for errors, omissions or inadequacies in the information contained herein or for interpretations thereof. The reader assumes sole responsibility for the selection of these materials to achieve its intended results. The opinions expressed herein are subject to change without notice.
Figure 1 Commercial ISV Support Profiled by Market Momentum Size of Application Portfolio IBM OS/400 IBM z/os IBM AIX Windows HP-UX Solaris/SPARC Safe for virtually all mainstream deployment Safe for many target segments and installed base leverage Limited verticals, niche segments and installed base only SCO Unixware Red Hat AS HP Tru64 Unix HP OpenVMS SuSE Solaris/x86 Other Linux HP NonStop Linux zseries Independent Software Vendor Enthusiasm Source: Gartner Research (September 2003) Figure 1 charts the life cycle of operating systems, which follows a counterclockwise path. Some operating systems achieve a huge software portfolio that subsides as they reach legacy status. Others only ever target a much smaller audience, but are successful within it. The top right corner of the chart indicates operating systems that have a large enough portfolio to satisfy users in most vertical and geographic markets. Most enterprises can safely base their IT infrastructures on these products, if the chosen product satisfies their scaling, availability and other needs. Only Windows and the top three RISC Unix variants (Sun Solaris on SPARC, Hewlett- Packard's HP-UX and IBM's AIX) fall into this category. The middle section indicates operating systems that fall into two categories: Systems in great demand, which may have a small software portfolio that will grow in the next few years. These products may still not reach the top right corner of the chart. Examples include the most successful commercial Linux implementations RedHat Advanced Server, SuSE and Linux on the IBM zseries. 17 September 2003 2
Well-established legacy operating systems that enjoy strong installed base support. ISV enthusiasm is likely to decay gradually as the ISVs search for new business elsewhere. Examples include IBM's OS/400 and z/os. Both types of operating system are likely to have a strong software portfolio, but are only appropriate for selected applications that match their characteristics. For instance, none of the Linux distributions is viable for companywide deployment yet. They do not have sufficient maturity throughout the software stack and their vertical scaling is not proven. But many verticalized applications will suit these products. The operating systems in the bottom left corner of the chart fall into one of three categories: Systems in great demand, but even newer products. Their software portfolio will be very limited in size and scope. Linux on the IBM zseries fell into this category in mid-2002. The entry for "Other Linux" would fit this category now. Even though Linux has great market momentum, the software portfolio for less well-established distributions remains limited to selected vertical and geographic segments. Highly successful in their chosen segments, but targeting a finite market need. HP's NonStop Kernel (NSK) is a good example. NonStop platforms have been sold as fault-tolerant systems for many years. NSK has achieved an enviable penetration of its chosen markets. But the market itself is small, and this limits the size of its software portfolio. HP is attempting to position NSK as a wider operating system choice by exploiting the trend toward the real-time enterprise. Niche operating systems that only ever address an installed-base market. These may be mature, declining fast or simply never gelled with the user community. HP's OpenVMS and Tru64 Unix fall into this category for different reasons, as do SCO's Unixware and Sun's Solaris on x86. We excluded Solaris on x86 from the chart in late 2001, but have reinstated it now that Sun has launched its own x86- based hardware and is attempting to position the OS as a Linux alternative in certain segments. We would only ever recommend the new deployment of operating systems in the gray band if there is really no other choice; for instance, if a necessary application runs only, or substantially better, on the operating system in question, or if price/performance advantages are compelling. Even then, we would caution users to validate ISVs' support plans carefully, and demand long-term support pledges in writing to ensure that the choice of such operating systems cannot be compromised. 17 September 2003 3
ISVs have been more enthusiastic about Windows NT on Intel than other platforms since 2000. But Windows 2000 Datacenter does not have as large a portfolio of certified applications and ISVs are less keen on it than on Windows 2000. Our model does not indicate each Windows operating system separately. "Windows" represents the portfolio for the Server and Advanced Server systems deployed by most Windows users. As Microsoft increases its investment in alternative Windows products, like embedded Windows and 64-bit Windows for Itanium, we will review this policy. Users that wish to run Windows Datacenter should check that their chosen ISVs actively support the Datacenter version. Signs to look for include implementations and technical skills in your region. Only applications that directly access the kernel need Windows Datacenter certification. Most applications written for other Windows variants should run and be supported by hardware and software vendors on Datacenter. Only servers that support more than eight processors, such as the Unisys ES7000 and IBM's xseries 440, can exploit the scaling assets of Windows Datacenter. Even then, many users deploying eightway platforms will often select Windows Advanced Server because it has a larger portfolio, costs less to deploy and staff with the right technical skills are easy to find. Hewlett-Packard's HP-UX and Sun Microsystems' Solaris (on SPARC only) continue to be the best platforms for vendors that support Unix. ISV support for IBM's AIX has increased recently and its portfolio is catching up with those for HP and Sun. Solaris used to receive as much support as Windows, but its position has weakened since 2001, leaving it only just ahead of the other Unix variants. First Windows, then Linux have become more popular with developers. The link between ISV enthusiasm and revenue can be demonstrated by measuring revenue, instead of ISV enthusiasm. The various operating systems do not move much (see Figure 2). 17 September 2003 4
Figure 2 Commercial ISV Support Profiled by Revenue in 2002 $15 Windows Solaris/SPARC $5 z/os HP-UX Revenue AIX Netware OS/400 $1 OpenVMS Linux/zSeries Tru64 Red Hat SCO SuSE Solaris/x86 Software Portfolio Source: Gartner Research (September 2003) Bottom Line: Use the independent software vendor support profile to gauge the probable lifetime of server operating systems. Only Windows and the top three Unix variants will suit those determined to deploy just one or two environments throughout the entire IT infrastructure. Operating systems targeted at horizontal markets may not be highly scalable or offer high availability, especially in challenging environments like batch, business intelligence or fault-tolerant computing. Other operating systems may address particular segments very well, but their software portfolio or life expectancy may be limited. New developments like Intel's Itanium 64-bit technology and planned advances in system partitioning and workload management software will give users the option to deploy multiple operating system versions and mixed-mode environments. 17 September 2003 5