Open Solaris Source: Customers Pleased/Competitors Vexed
The recent announcements made by high-ranking members of Sun Microsystems with regard to an open-source version of the Solaris operating system has created quite a stir. It would be wise to remember the history prior to the recent attempted rapprochement between Sun and arch nemesis Microsoft, before understanding the true relevance of these announcements. At first glance it might appear that Microsoft and Sun have the most to lose in the face of the widespread and increasing adoption of the Linux operating system. The latest 2.6 version of Linux is able to support the demanding needs of high-volume mission-critical applications that were once the exclusive domains of the established corporately supported UNIX operating systems (e.g., Hewlett Packard HP-UX, IBM AIX, Sun Solaris, etc.). Now, Microsoft not only faces continued pressure at the enterprise server level from Linux, but also upcoming Linux desktop offerings that provide enhanced features and functionality that are combined with improved user interfaces. The continued discovery of new security vulnerabilities exemplified by last year's Blaster worm have stifled corporate sales and exacerbate that situation more so when coupled to corporate resistance to Microsoft's attempted transition to subscription-based licensing with its annual or biannual fee structure. These security concerns keep open the door for both server and desktop Linux alternatives. As a result, the April settlement between these long-time rivals should come as no surprise. Open-source Solaris is just the first major salvo against Linux from Sun, which is likely to be followed a massive Microsoft offensive. During the next 12 to 18 months, expect the Linux wars to intensify. It will be very interesting to see what the rationale for bolstering the established operating system and application software branding will be.
Senior members at Sun made it clear in late June that an open-source version of Solaris is in the cards, but not open-source Java. Case in point: Jonathan Schwartz, president, and John Loiaono, EVP of Software, have stated that Sun will offer an open-source version of its Solaris operating system. Ironically enough, around the same time, Sun CEO Scott McNealy made a statement strongly rejecting the idea of open-source Java. Making matters more interesting is the apparent contradiction from executive director of JES Marketing Stephen Borcich on July 15th that the company was "considering open-sourcing the Java Enterprise System (JES)." The JES package includes the firms' application server and directory server, along with collaboration, communication and security offerings. In a nutshell, the IT community can realistically expect an open-source version of Solaris by the end of this year, but open-source Java is anyone's guess.
Sun: Still a Powerful Force
Prognosticators' claims that Sun has become a "has been" or "scavenger" reflect a suspicious origin, since it has no basis in reality within the top-tier technology giant and should be dismissed as simply irrelevant sound bites. While it has appeared somewhat short of a singular guiding vision in recent years, Sun is still a powerful leader in three areas of IT that include 1) Standards: The Java Community Process (JCP) is the primary mechanism for the open, participative process to develop and revise the Java technology specifications, reference implementations and test suites; Sun's Liberty Alliance provides a second public voice in the ongoing effort to arrive at an industry-wide single sign-on option for basic user authentication; and, Sun regularly participates in W3C and OASIS; 2) Hardware: Offerings include Sun Fire E6900, Sun Fire 12K and Sun Fire 15K; 3) Software: Offerings include Sun Java System Application Server, Trusted Solaris Operating System and Java Studio Creator. In addition, Sun announced at JavaOne in late June its plans for service-oriented architecture (SOA) and Project Looking Glass desktop windowing system.
A New Business Model
With open-source Solaris, Sun creates a new business model for the enterprise. Linux's biggest selling point has been the cost advantage over cost-type UNIX operating systems. Open-source Solaris ends that advantage in one fell swoop. For example, a financial services company order management system (OMS) currently runs the Solaris 9 operating system on the mid tier (Oracle9iAS and Sun 6800) and the back end (Oracle 9 database and Sun 12K). This hardware/software setup provides excellent availability, reliability, scalability and security. If the company wanted to increase overall performance, a variety of options such as Oracle real application clusters (RAC) and parsing options that include Simple API for XML (SAX) and Streaming API for XML (StAX) are currently available. Upgrading the application server and database to the latest Oracle 10g versions could also increase performance to desired levels, while still retaining the preexisting Sun Solaris 9 operating system and mid-tier and back-end Sun hardware. Open-source Solaris only makes things more attractive. How? The financial services company could 1) Leverage its in-house knowledge of the Sun Solaris operating system, Sun hardware platforms and Oracle software; 2) Retain its current Oracle application server and database software; 3) Retain its Sun mid-tier and back-end hardware platforms; 4) Lower costs by adopting open-source Solaris; 5) Leverage new Oracle and Sun development tools. These are five valid and compelling reasons, why the financial services company would choose not to move over to the Linux operating system.
However, while these developments provide a ready migration path for maintaining investments in current technology by leveraging the possibilities inherent in a strategy based on extending the life of Solaris through an open-source approach, it remains to be seen if developers accustomed to simply working within the current Sun support system will now be excited by the possibilities inherent in open source. Open source requires an almost familial loyalty and willingness to endure the growing pains of endless cycles of snapshot releases and overnight bug reports, all of which represent a drain on Sun and client resources if they join together in Sun's Open-Source Solaris project.
That is no small question, and while Sun's JCP has succeeded in keeping Java at the forefront of application development and now enterprise architecture (EA) and SOA integration development, Java is a programming language not an operating system. An operating system represents an order of magnitude increase in the level of support needed. Simply hoping that something along the lines of JCP will work for Solaris would be an error. An effort of greater scope is required, and more of an enticement than simply losing the cycles of upgrade costs for the operating system needs to be developed for clients to adopt the kind of effort required by open sourcing. There is a different cost and a different culture required to make open source succeed with an existing product, whose development has not been nurtured by a community of adherents that verge on zealotry as with Linux. There is considerably more buy-in required, not to mention the feedback that truly amounts to a kind of reward through shared suffering that it is doubtful that Sun's Solaris effort can mount. Unlike Linux or its own Java, Solaris doesn't come close to the kind of esprit that maintains those efforts. Oddly, the one avenue that might provide that kind of religiosity would be open sourcing Java itself. In that case, the investment that a large community has made in JCP could "rub off" on Solaris, provided there were integration advantages that could be built into Solaris for Java. Then the independent developer communities would have a much larger incentive to take on Solaris, relieving the installed customer base from the necessity of allocating resources to that effort. As noted, it will be interesting to see what develops.
Execution, Execution and Execution
If carefully rolled out with explicit advertising in trade publications aimed at systems integrators and network engineers, explanatory articles in IT publications aimed at industry decision-makers and consciously encouraged word of mouth, open-source Solaris can slow down the attrition from Solaris to Linux-supported operating system offerings (e.g., Red Hat, Novell (SuSE), TurboLinux, etc.) and from SPARC to Linux supported hardware offerings (AMD and Intel). However, as noted previously, this must be coupled to the kind of advantageous support for Java that Microsoft affords to Visual Studio .Net while simultaneously delivering the rapid pace of assimilation for improvements and bug fixes seen in the Linux community. This is achievable, but requires follow-through and corporate commitment to supporting the open source Solaris community that it must first encourage.
In addition to more stable hardware revenues, the firm will have a greater chance of selling software products such as application servers, portals and development tools to pre-existing customers. Linux-promoting hardware vendors such as Dell, Hewlett Packard and IBM will certainly be keeping a watchful eye on open-source Solaris because they have the least to gain and the most to lose from this latest Sun effort. If these competitors can be provided with good reasons to buy into this effort rather than fighting it, Sun will relieve itself of one of the more onerous tasks in platform-level software/hardware development, gaining adoption. It might be wise to give thought to pressing this issue as a benefit for all rather than a raid on competitors' clientele. In essence, the cost-type UNIX versus non-cost type UNIX debate could effectively end with open-source Solaris.
This is both a stabilization and offensive move on the part of Sun. As with most serious opportunities, open-source Solaris can also be a minefield of potential difficulties, obstacles and hurdles. It all depends on how it is approached and supported. Merely waving their hands over a source code release and hoping that it will be picked up and carried forward by hordes of salivating software developers would be a prescription for more red ink than could possibly be saved in terms of reducing the cost of further Solaris development.
What Should We Expect?
Rumors and gossip aside, what should we expect in the way of a Solaris offering during the next 6 to 12 months? Sun would most likely field three offerings that include:
- Full Featured and Priced. Encompasses all the latest features and functions, along with enterprise support.
- Full Featured Open Source. Encompasses all the latest features and functions, but with no support.
- Full Featured Open Source with Blocks of Support. Encompasses all the latest features and functions, with the addition of blocks of support that may be purchased by businesses and organizations on a when needed basis.
Number three should be by far the most popular option because it will allow Sun customers to purchase blocks of support (at a defined cost) when needed. As with Linux, no open-source software (Apache, JBoss or MySQL) offers a "free ride." Businesses and organizations often require support, and such in-house developers using open-source Solaris can take comfort in knowing that Sun will always be there if the need arises. Outsourcing developers, integrators and network designers will benefit by having the option to specialize in customized adaptations of a well-known platform with proven performance at scale. Add viable hooks for easier Java integration and you have a very attractive package.
From an engineering standpoint, Sun makes some of the world's best, most reliable, hardware and software products. In addition, it must be considered the leader regarding Java standards with JCP. Make no mistake; Sun was a market leader before this year's settlement, which has provided the firm with a nice chunk of cash (approximately $2 billion) and a very powerful potential partner in Redmond. Open-source Solaris must be viewed as both a stabilization and offensive move on the part of Sun to slow down the Linux juggernaut. Dell, Hewlett Packard and IBM will be keeping a close eye on this first post-settlement Sun foray and new offensives from Microsoft.
Is this the opening gambit in the long overdue unifying vision that Sun has lacked in recent years? If so, then Sun and Mr. McNealy have a wonderful opportunity to increase revenues and regain market share.
For more information on related topics visit the following related portals...
Russell Ruggiero is a senior IT analyst. He is the acting chairman of HumanMarkup.org. Ruggiero has authored more than 150 articles and reports for well-respected firms that include Gartner, Inc. and Source Media. He may be reached at firstname.lastname@example.org.
Rex Brooks, president of Stabourne Communications Design, has pursued an extensive and wide-ranging career in advertising art direction, corporate identity and graphic design. His ongoing interests have included the applications of computer technology in his field and applying concepts from the fields of psychology, sociology and advertising in the area of semantics and semiotics for the purposes of improving communications in digital information systems. This led to his involvement with OASIS in the HumanMarkup Technical Committee helping to create the Human Markup Language. He is the cofounder of the Content Development Working Group of the Web 3D Consortium and Humanmarkup.org, Inc. and serves as vice chair of the OASIS HumanMarkup Technical Committee. He is also actively serving on the OASIS Web Services for Remote Portlets and Emergency Management Technical Committees. You can reach him at email@example.com.