Last week, Sun and Fujitsu announced the convergence of their Enterprise 64-bit platforms. They will co-develop and separately market a SPARC-based Advanced Product Line (APL), available 2H06. This is a good step that combines previously fragmented competing server product sets - Sun’s UltraSPARC-based SunFire systems and Fujitsu’s SPARC64-based Primepower systems. The move strengthens the overall SPARC position in the data center as a viable alternative to IBM Power and HP/Itanium Unix platforms.
This will be an interesting exercise for both companies in an engineering and commercial sense. Commercially, where Sun is dominant, this is a complementary strategy. In Germany and Japan where Sun and Fujitsu compete, or in growth markets for Fujitsu (e.g., US, UK, Scandinavia), product positioning and overlap will be a challenge, and competition will continue between the companies - without a commercial extension to this engineering agreement. Indeed, we believe this convergence of products will open opportunities for service contracts from one vendor to the other in some expanding markets.
Although the announcement states that APL will be an end-to-end hardware platform, we believe the engineering efforts will help position Solaris SPARC as a back-end platform provisioning services. Moreover, this will help leverage the mainframe heritage and self-healing features of Fujitsu and may help finally overcome any reliability issues that have arisen on Sun’s platforms. This will be appealing both in the data center and in general-purpose server platforms.
Fujitsu will move from being a Solaris-compatible platform to the same platform as Sun, and this will strengthen its position in the marketplace and against Sun. Fujitsu will also use this platform to consolidate its mainframe and high-end Unix business, as well as support the platform with Lintel and Wintel Web and application servers on reliable BEA middleware and EMC storage services - though Fujitsu will also support Java Enterprise Server (JES) and Java Virtual Machine (JVM) within Solaris 10 as part of this agreement.
We also believe Fujitsu’s efforts in heterogeneous platform management will help control and deliver these mixed-platform solutions across the IT organization’s portfolio. Indeed, Sun must make similar efforts to offer common management tools and control facilities to all its supported platform mixes, Solaris/SPARC and Intel as well as its recent Lintel and Wintel platforms. However, Fujitsu does have to carefully position its Itanium servers against these Solaris platforms.
We see this as a step on the path of change from Sun, and as a reflection of the recent changes made in the management team with Jonathan Schwartz. These changes include delivery of Lintel/Wintel and AMD and Intel x86 servers, alongside radical “emerging economy based” licensing models for software, the JES, and the Java Desktop System (JDS) additionally focusing on government. Moreover, Schwartz has indicated (but not stated when) Solaris will become open source, proving that Sun is aiming to move to a software and services revenue model to offset its declining hardware margin. This SPARC platform agreement with Fujitsu proves Sun is listening and acting on required changes. It now needs to deliver on these messages and offerings commercially.
Bottom Line: Through 2009, Fujitsu and Sun customers should position x86 Linux and Windows as well as Solaris SPARC Unix platforms in the portfolio (alongside HP-UX/Itanium and AIX/Power platforms). Any other mix of operating system and processor should be classed as niche and either validated or rationalized accordingly.
META Group originally published this article on 8 June 2004.