Sun software noise
Is it too premature to ask whether Jonathan has failed? Let me come back to this...Sun has made three big moves in 2008, outside of the non-stop nonsense around openSolaris: MySQL, Glassfish, and JavaFX; let's break them down, since I have some hours to kill...
MySQL at $1B is a laugher, that much can be agreed upon by everyone except Sun's investor relations department...i understand that it is symbolic of the extent to which OSS has penetrated the enterprise market, and that it probably took a significant price to close the deal without a bidding war, but the execution of making that investment bear out will be as difficult as anything that Sun has ever undertaken...truly, the only measurement that should be taken in to consideration of whether the acquisition was a wise one, is the ability to migrate Oracle DB customers on Sun hardware over to MySQL, not whether some Web 2.0 companies expand their services relationship with Sun...
Oracle with the acquisition of BEA, its ongoing efforts around 'Unbreakable Linux', and its pricing schemes around multi-core chips are all the proof needed to finally put to rest the supposedly unmatched relationship from these Highway 101 legends: Chuck Phillips runs Oracle's strategy and Jonathan runs Sun; the historical "friendship" between Larry and Scott is no longer a factor, and Oracle has more to gain from Intel than Sun, regardless of the press releases and whatever the two companies' sales forces tell customers...Sun needs to port SQL environments to MySQL off of Oracle to maintain account control, at the same time that Oracle tries to maintain the WebLogic-bias in Sun GSO; how Sun makes MySQL competitive with the unquestioned leader in the database market remains to be seen...i am not buying anything that Jonathan says ab/ MySQL focusing on cloud initiatives, because if this were true, there is absolutely zero way to justify $1B...
Glassfish is just definitively saving the Sun software business, from the SeeBeyond debacle, to NetBeans' failure, to even Solaris' slow death; its Teleco implementation, its ESB, and the ISV opportunities, such as with Liferay, are giving a life-line to Sun GSO, if they could only just make the transition to selling away from WebLogic and on to Glassfish in the downtime while Oracle figures out Fusion, there is hope for all the die-hards (including myself) to justify the immense investment in enterprise software initiated with the Sun-Netscape Alliance...i will admit my historical and biographical bias, but nothing comes close to Glassfish in providing Sun some hope and a raison d'etre for remaining a viable systems vendor...
JavaFX is plain ridiculous, wrong-sighted, and D.O.A., enough said...
Now back to my former big boss, as Jonathan was Sun's first EVP of Software prior to his anointment as CEO; his accomplishments are well-known by Sun insiders, most notably Java Studio Creator, Orion, and the Liberty Alliance...what is the common theme of those 3 initiatives? they are undeniable failures, which calls in to question his grasp of software, fundamentally...Creator was E.O.L.'ed, Orion is not even relevant to any product at Sun, and Liberty is a joke to anyone who has even heard of it; combine this with openSolaris, which is not viable in the face of Linux (as anyone outside of Sun would tell u), and u have to wonder what he is doing to make Sun an alternative to IBM and HP services, the Oracle stack, and Dell's model...
I'll admit I remain a long-term buy on JAVA, but I am increasingly concerned that the small circle of executives with a say, let alone the apologists on the Board and among major shareholders have either no grasp of the emerging enterprise IT marketplace or are just too timid to keep Jonathan in check...I'll also admit that there are possibilities he will prove me wrong, not least because the guy is relatively young and has enough cash, goodwill, and employee talent to ride out the ongoing disappointment that is quarterly reports...
But if a spade were truly called a spade, I would say that McNealy's hope to turn Jonathan in to a Steve Jobs-like wonderkid with the transition have not been borne out: revenue and the stock are stagnant, the Solaris-mindset is categorically narrowly-focused to the point of being a death knell for the company, and without talking with anyone internally, i sense that employee morale is weakening, as the 10% policy, as well as ongoing additional cuts, take their toll on execution...
I am a very small minority, or at least a minority in the sense of those willing to speak the unspeakable, but i will close with the question for Sun stakeholders to ask: how much longer? He is a first-time CEO (still unclear to me how Lighthouse was considered a business worth buying, let alone existing) of a Fortune 500 company...and i am not suggesting that the Red Hat model of bringing in a seasoned outsider is more viable option, so i am not sure what to do...however, giving a newbie a long leash is one thing, running an American icon in to the ground is quite another...
I'll invariably lose some friends/contacts over this post, but would be more than willing to hear from those that are capable of refuting the contents i mention above, I wish all of Sun the best in its endeavor to recover from wounds that were, for-the-most-part, not of its own doing in 2000-01, but it is 2008, its major competitors are growing and moving on, why can Sun not do the same? I am open to your suggestions...
is JBoss dying?
I probably only have a brief few comments to make as I don't believe the implication in the title of this blog entry, but JBoss is certainly facing an evolving competitive front, and basically has not responded to it, unless u count the diatribes of Bill Burke, which essentially leave me wanting to defend, and failing to come up with the motivation. The competitive fronts are two-fold, and will be recognizable by even the most passing of observers in the middleware market:
1. Seam is losing month-by-month to Spring.
2. JBoss is losing month-by-month to Glassfish.
Ok, ok, I know for anyone who has had to listen to my own diatribes on TSS, on blogs.sun.com, on Maison Fleury, or any other forum that Dave Rosenberg does not control (and sometimes even there), this two point formula is well established to the point of ad nauseum banter from me. It also fails to mention the competitive front from Oracle's acquisition of BEA, which is a seismic development, but ultimately fails to seriously threaten the long-term viability of JBoss: Spring and Glassfish, however, do...
Take a look at the corresponding proof-points, first the JCP page for Web Beans, the standardization process of Seam:
http://jcp.org/en/jsr/detail?id=299It is stagnant, while Spring rolls out S2AP and a myriad other complimentary technologies to seize the developer's attention. Seam is dying on the vine, and its own complexity does not bode well for a major roll-out of improvements any time soon. This is coupled with data point #2, which is JBoss' embarrassing inability to finalize and release JBoss 5:
http://www.jboss.org/jbossas/What on Planet Earth is going on at JBoss, is it true, according to Marcf, that the middleware division at Red Hat, as represented by all of the various JBoss products and projects is under-resourced? Or are we to believe that Exadel, the SOA product-line, and that silly Tuxedo-replacement project is as important as the core Java run-time? I have been wrong before, but i am not buying it...
I am starting to question the ability of new management to take over JBoss' model, even if Red Hat is OSS, it is starting to remind me of the slow but undeniable slide under Alfred after Bill Coleman. Not that Alfred was to blame, there was (as apparently with JBoss) mission-creep in the form of Portal, Integration, and the proprietary Java web services technology that locked developers in to Workshop...But just look at jboss.org, it is a mind-numbing array of ongoing and outstanding projects, with SSO, ESB, and Rules among others seemingly having no delivery date in sight, that translates to project stability...
Please, all u JBossers, come on here and tell me how i'm wrong, i am a well established fan of the history, and hopeful proponent of the future, but things look awry...the Java community needs JBoss, even if Spring is answering some of the development model requirements that were once the sole domain of JBoss, and Glassfish is just going from one impressive enhancement/announcement to the next, but JEE5 and 6 need a representative JBoss app server, otherwise customers will have to look to migrate, its that simple...
I am on the record stating a vision of portability for developers that leads to inter-operability of components across the business environment, that does not come from WebLogic, WebSphere, or even Spring, necessarily...it comes from JBoss, the continued market leader, that seems to be weakened, and if momentum were a gauge of leadership, to have lost the mantle that it so perfectly executed to acquire...what is going to come next that translates in to a reversal of the current state of things in Java-land...
I honestly don't know, but a few recommendations come to mind:
- skip JBoss 5 in favor of 6 (scary thought)
- pare down Seam to a JSF model and leave EJBs for a future rev. (probably too far along to do this)
- outsource some functionality that has been supposedly brought in-house, such as the SSO, ESB, and Rules mentioned above (requires a major policy shift on the JBoss management plan)
- go Linux only (this seems to be the most viable); sure u can have a Windows development environment, but honestly is it strategic to have a deployment environment when 90% of Microsoft shops are going to be doing .Net, anyways...
Its time to start answering the hard questions, and new management is no excuse, perhaps going the tried-and-true route of brining in some airline exec. instead of Marcf was a mistake that too heavily favored the OS business over the ever-more lucrative middleware market...i am just thinking out loud here, and have pissed off too many people in the industry to care whether this hits the JBoss crew as inflammatory: i want to see JBoss succeed, right now, it is definitively not...
Spring arrives
(I have not been on this blog site for some time, but did write about
Glassfish on another blog I maintain; this is my response to the latest news from Spring)....
What is Spring?
In its most basic form, it is a development paradigm, disguised as a component model. Its value proposition is 2-fold, one stated, one not: create higher developer productivity and the unstated goal of replacing EJB. Through a transformation from an open source, proprietary development framework for Java to a borrowed-source, standards-based deployment run-time for its own software, SpringSource is attempting to become a ubiquitous presence in the decision making process of enterprise IT development staff, and potentially in to the echelons of IT management. In short, no company has come as close to the heady achievements of JBoss, as has Spring and its corporate sponsors.
This is why VC's have put money to the cause, and believe that they can achieve the sort of returns that have guided previous calculations on JBoss and MySQL. But even beyond exit strategies, Spring challenges the very foundation of Java and its corporate sponsor, Sun Microsystems with this attempt at ubiquity. For beyond the hype of JSE and JME for desktop apps and mobile devices, respectively, is the true raison d'etre in JEE, where all the Java money resides. Embracing and extending JEE is something that not even Microsoft has figured out, so it is with optimism tempered with caution that the JEE developer utilizes Spring.
Of course, for many, the question is: why caution? By decree from supporters, Spring has become a "de-facto standard", and should thus enjoy the same support and confidence as EJB, in its role as component model for the JCP-managed JEE specification. This is debatable, in the absence of true ubiquity, and the debate has carried on web forums, and presumably within customer accounts. Caution is reserved for EJB as well as Spring in the determination of what constitutes a standard. And ultimately the determination is an academic one, for IT staffs have varying degrees of risk vis-a-vis a standard, let alone different qualifications for what makes a so-called standard important. Suffice to say, when all is said and done about "legacy" EJB and "de-facto" Spring, they are roughly on common ground when claiming standards supremacy.
What is important about being a standard of any kind comes down to a single concept: portability. This takes two forms, along the lines of the component models of EJB and Spring. For EJB, it is through the Application Verification Kit (AVK), and for Spring components, it is assured through the presence of the Spring Framework in the environment which the components run. Both have their strengths and limitations, but without exploring in much detail, and similar to the standards claims of both technologies, EJB and Spring, it is suffice to say that user beware of the portability value proposition.
In short, EJB and Spring, controlled by two very different, yet competitive entities, in Sun and SpringSource, face a future of controlled confrontation. Sure, SpringSource will sit in on Sun's JCP efforts to standardize JEE6, but this will be at expense of unanimous support from vendors on its component model, in the form of EJB 3.1. There are a myriad of sub-factors involved in the emergence of a Spring-endorsed JEE6 specification, from profiles to modularity, but the ultimate outcome is almost assuredly going to be ongoing competition between Sun and SpringSource, between Glassfish and S2AF, and between EJB and Spring, itself. Not even JBoss is relevant enough to engage in the discussion, even with Seam.
And so it will be an incredible display of marketing and engineering in genuity to upstage the other, occassionally with the support of a complimentary vendor for EJB (in the case of JBoss), and for Spring (in the case of WebLogic, assuming its longevity within Oracle).
But ultimately this is an execution based market where the spoils are the ever-monetizable quality of enterprise developer support. Sun and SpringSource have a lot at stake, coming from different stages and both attempting to grab the mantle of counter-weight to .Net. Who will win? Will it be the improving, but tainted EJB? Or will it be the upstart, yet relatively new Spring and its cadre of vocal supporters? Ultimately, only time will tell, and any bet is as good as the opposing one, but the one agreed upon statement is that whoever or whatever bests the other, EJB or Spring, a great deal of IT money will follow. A stalemate of equal but separate marketshare is not an option.
