Astro MicrosystemsAt first glance, evaluations of enterprise systems vendors are a prime example of analysis of a vertically-integrated business operation. Sometimes it comes in the form of a farmer, or farm output, such as potatoes, where the potato farmer owns the land, the mills, the specialized fertilizer, and the transport to carry the product to market. This is a long-discounted practice, as specialization offers comparative advantage and outsourcing benefits. In fact, outside of some financial services providers and Information technology vendors, vertically-integrated businesses are being replaced by overlapping business interests that promote the Wall Street requirement of focusing on 'core competencies'. Even financial services firms are probably on the way to being subjected to new regulations that will greatly reduce their scope in providing a 'one-stop-shop'.
The same can also be said for IT firms, but not necessarily because of regulations. Microsoft, Red Hat, Intel, Dell, HP, and Oracle all rely on some partners to facilitate their highly successful businesses. The lone hold-outs appear to be IBM and Sun, and perhaps some Asian chaebols . However, IBM has relied on more partners than ever to deliver the returns that primarily come from IBM Global Services, their consulting operation, which generates the bulk of their revenue, growth, and even profits. Asia has a historic relationship with vertical integration, and will probably reduce their exposure to inefficiencies in some aspects of their business by partnering with global firms to help facilitate their emergence in the global economy. That leaves Sun. And it is to Sun that I turn my attention (per usual).
Sun, in fact, remains a huge company with $13 billion in revenues and 35,000 employees worldwide. It produces everything that it sells, for the most part, from microprocessors, to hardware servers, to storage systems, to software servers that run the gamut of functionality that enterprise customers need to run their on-line and back-office operations, outside of ERP. SPARC, Solaris, StorageTek , and Java are their core businesses that represent the vertically-integrated business that continues to buck the trend toward specialization, and there is a case to be made for continuing this verticalization . Sun is essentially a 'proprietary' company though they cover their bases with licensing Solaris to distributors that double as competitors. They also do an increasing amount of open source work, which is essentially the subject of this post.
By making their new chip research available through openSPARC, their storage systems available through openStorage, their operating system available through openSolaris, and Java available through Glassfish, among other projects such as openJDK and MySQL, Sun has burnished an image as a community builder and a purveyor of standard components that make up an IT system. Sun is essentially making their core assets open source to generate developer interest, open up OEM opportunities, and create customer confidence that though a purchase of Sun gear may be a proprietary investment, there is a channel to maintain their investment in the off-chance that Sun ceases to exist. It is a reasonable bet, and careful scrutiny of the IT marketplace results in basically a sign-off on the strategy. Sun, were it to grow these various "open" communities that represent their various product development channels, could be considered well-positioned to take advantage of the Internet-fueled, next-generation of customer deployments, environments, and implementations of the IT systems that run the world's businesses.
But if you are reading this still, you are aware that this scenario of Sun's impending success is, in fact, no guarantee, nor, as many would contend, at all likely. Why is this? Well, without getting in to a big debate about the likelihood of Sun's ultimate end-game success or lack thereof, I rather point to serious concerns for investors, for customers invested in Sun, for developers relying on Sun for technologies, and ultimately the entire IT ecosystem. In reality, if there were one tier-1 vendor that could in-a-flash disappear without causing much damage to the global economy, it would not be Red Hat, certainly not Microsoft, Intel, IBM, or Oracle. It would be Sun. The continuing inability for the company to grow anywhere near its closest competitors, and the fact that perception matters in IT decisions, means that the sudden disappearance of SunMicrosystems is a real possibility.
And with that background, I would like to propose a check on the downside ramification in this scenario, by proposing an alternative to a Sun-only world for Sun technologies, by creating a competitor based explicitly on the R&D output of this venerable IT firm. Essentially, as I have alluded to in previous passing comments on tech news pages, I am proposing a new company to be created out of the open source communities that Sun has created. In the relevant parlance, we call this 'forking'. The term forking has a typically negative connotation, but this is just because of the incompatibilities created for the community to keep up with which results in less than 100% focus on the set of problems and challenges that are involved in open source. But the new company based around the open communities Sun has engendered would not seek to create incompatibilities, in fact, considering the relatively impressive customer base using Sun technologies, it would be in the best interests of this new company to maintain widespread and ongoing compatibility with every output from Sun.
The basic construct of this new company would simply be the development of a platform based on Sun's most relevant, most promising, and most attainable technologies, all of which are freely available through their open communities. At this point of the diatribe, I will point out that there is precedent, in one huge, though as yet under-developed, example in Oracle's Unbreakable Linux. In fact, my exit strategy for this new company may be even to receive the support, funding, or acquisition by Oracle, but that is just a glimmer-in-the-eye idea; rather, I take a cue from Oracle's efforts to take Red Hat's Linux operating system and turn it in to one of its own products. For those familiar with that matter, is it really that much of a stretch to do what Oracle has done with Red Hat to do with Sun's open technologies. Granted, much difficult work would have to be undertaken to create hardware systems as well as software platforms, but there is nothing that says that it could not be done. And, frankly, I cannot, in all my travels, and somewhat limited experience in the business world, think of a start-up opportunity with more up-side potential than a Sun Microsystems-clone.
First, take Linux, license it from Red Hat and/or Oracle. Next assemble Azul engineers with expertise in multi-core chips built for Java middleware to take the specs from openSPARC and create a computing platform. Then, utilize the vast Java developer base to further the assembly of Glassfish and MySQL environments. Sun has a run-rate of about $12 billion in R&D, COGS, and SGA, which I think could be cut by 75% with this new company, considering the vast amount of inefficiencies that have grown around Sun like plaque on teeth. Undercut Sun's own product-line by 50%, and you do the math on profits and profit potential if you really believe any of the hype coming out of Sun that they are ushering in the next new utility switch for the Internet age. I will make one prediction around this scenario: it is only a matter of time.
Sun has been an innovator of computing systems and has been a fairly advanced operational company, at times. It has, however, achieved none of the potential afforded it with so many valuable assets. It is now time to challenge the conventional wisdom that what holds it back is lack of understanding by the IT customer base, or the whispered sentiment coming out ofMenlo Park, CA that the world just doesn't 'get it' yet. It is time to call a spade, a spade, and realize that Sun Microsystems ' best years ended in 2000, and there is no discernible path to profitability in its current form. It is time, at least, for competition to inject a sense of urgency in to Sun. But this is no charity cause for the re-emergence of a once-hallmark of American ingenuity: no, this is about taking money away from Sun, and doing it on a truly huge scale. I will refrain from critiquing management, but from my experience, I can attest that there is a virtual "government-in-exile" waiting to take up the leadership and middle management positions of this proposed venture who know Sun very well. All it takes is money and courage.
The technology is there, take a night of surfing, and look for yourself, and realize that potentially the biggest investment opportunity since the great advent of Netscape, of Google, of Sun itself in the 1980's is there for the taking. I am not ready to be CEO of this new company, but I would be a pretty good person to present this business case, especially around the long-forgotten competitive advantage of Sun's in the form of Reference Architectures. Building an appliance with the following is all I propose; give it some thought, and you know how to get in touch with me:
use Linux, deploy Glassfish with MySQL on top of openSPARC with the 5120:
throw-in openStorage with the 4500:
Let's just call it Astro Microsystems...