the state of Hardware servers
O.k., Jonathan, apologies for the obsession, and apologies for discounting your analysis on the U.S. server market, as Dell reported weak server numbers at 5% growth for the
quarter...
Combine this with IBM's most recent server
numbers which grew a rather paltry 10 percent (4 percent, adjusting for currency), and perhaps i indulged in
hyperbole following the Gartner
report...
So, should i get on-board and accept that the U.S. server market is dead for the time being, and that the growth will have to come from BRIC and other emerging countries?
Should I issue a mea culpa and start taking expertise from Sun's CEO on the state of the Information Technology marketplace, and stop harassing him?
Perhaps...but, not yet, and I promise not to back-down from the openSolaris critique, this is something that I have assessed over a period of about six years, ever since Solaris 8U6, and still feel that Red Hat is the
buoy that will save Sun...
But what to make of the hardware market, absent some all-encompassing analysis from
Ashlee, who must be on vacation this week...:
I don't know, i still believe in the promise of multi-core, mixed with some middleware pre-installed and pre-configured for vertical solutions and/or high-performance computing requirements, whether that is "cloud" or not...essentially, I feel the Glassfish/MySQL
combo on the T2 will be Sun's only competitive advantage, even considering the
openStorage push that is about to commence from Santa Clara...
Dell will sell the "cloud" well, and IBM will grow on the backs of administrators worldwide tied to mainframes, or those who simply cannot manage their IT departments by themselves...HP will essentially pick-up the scraps (though significant scraps) from Dell and IBM in both areas...Sun, on the other hand, just doesn't have a choice but to fight, get, and stay out in front on so-called innovation, even as this thing we call a hardware market goes through intense change and margin reduction, accompanied with Unix-replacement by Linux, Vista-maturity, and Oracle consolidation...in an election year, coupled with international geo-political/economic instability, and basically little competitive differentiation, there is not much the vendors can do to demonstrate a business model for leadership...
so, Sun, you may appease the base investor community (in other words, JAVA may stabilize at its 52-week low, and not retreat much beyond that) with claims of maintaining 10% market-share as a success, but I am not going to take it lying down...absent
Rock processors coming out anytime sooner, T2 and the upcoming
T3 will have to hold the fort...this is doable, though scary for the faint-of-heart, especially as the cash is evaporating, not-to-mention goodwill, momentum, installed base, and even potentially a raison d'etre...
Glassfish is simply dominating the Enterprise Java market, as JBoss 5
falters and lacks clarity in its actual purpose and timetable, WebLogic tries to become Fusion, and IBM evaluates whether to even support JEE5 and 6 with WebSphere, proper (outside of Geronimo noise)....MySQL is a nice thing to have, and I am not going to do any second-guessing as long as the target is enterprise deployments, and not any unnecessary Web 2.0 banter...so, as i have said many times the Niagara hardware with T2 and T3 could be enough with the middleware, if Reference Architectures were resurrected, and if Red Hat were supported...but instead of such focus we get the openSolaris nonsense, masked in openStorage marketing, supplemented by an openSPARC inspired
OEM division that is a non-starter...
I know it is infinitely easier to take pot-shots than run a Fortune 500 company, but i am trying not to be 100% inflammatory, for I do want to see change more than demise or eventual failure, I don't want Jonathan's job (i am not moving back to CA), I just want a way to piggy-back on some badly needed success, and make some money off of Glassfish, so i understand that the challenge is almost over-whelming, but there is a sliver of a chance that things could turn around, especially by 2009; i know Q1 is going to be brutal, and I know i will host another session of criticizing the public statements of the CEO and CFO, especially if it is even close to the last
call...but I am willing to bury the hatchet, if Sun can do the same, i am offering an olive branch, but will fight with vigor if the response remains the same...
Sun is well positioned, and the pot of gold may be within reach, contrary to my headline quotes from previous entries, but its not going to be easy and its not going to come from simple inertia, its going to come from tough decisions, and a change of emphasis, and more of what I have been saying all summer...I am not claiming to be infallible or the only one possessing the right answers, but i feel pretty comfortable that i am right about a good deal of it, and no one has contested some of the major provisions of the turn-around plan, if paralysis has set-in, and things don't change, "Sun Alumni" will just be another way of explaining a Silicon Valley truism: almost everything goes away eventually...
openStorage
I am relieved to have been distracted the past week (Usain Bolt, people), but I am keeping my eyes open, and have a simple request to the few stragglers at Sun that glance at my blog either through a hidden bookmark, or on JavaWorld's page:
get Jonathan to read my blog...
I know, self-aggrandizement and self-promotion are two qualities that do not go well together, but every Monday from here until the Q109 results come out, I am going to cringe looking at finance.google.com and searching for JAVA news and stock-price: I mean, when is the investment community going to give-up? apparently the Sun executive team and the compliant PR department don't think as soon as most everyone else, as they allowed this press release to come out:
http://www.sun.com/aboutsun/pr/2008-08/sunflash.20080820.1.xmlNow, I know openStorage is going to feature more prominently in marketing, as indicated by PTB in the FY08 con. call, but it is just a new way of saying the same old crap about openSolaris, something that no one - - and I literally mean no one - - is paying attention to...I can't even bring myself to analyze this release of old news, as it is all well-documented in my previous
posts...
Something I will talk about is server marketshare, as the most recent update from Gartner came out last week:
http://www.theregister.co.uk/2008/08/22/gartner_server_q2_2008/No, actually, I am not going to do much justice, because i am just going to get emotional and inflammatory if I do go in to it, as readers of this blog will no doubt understand that my take on Jonathan's and Lehman's assessment of a struggling server marketplace is a flat-out lie, and categorically a mis-leading analysis...
go ahead, read the article on El Reg, and justify to me the statements of the CEO and CFO that the economy, in general, and the server customer-base, more specifically, is contracting: not to Dell, IBM, and HP...what do those three have in common: secondary support for Solaris...what do the two shrinking vendors - - Sun and Fujitsu - - have in common: primary support for Solaris...
So, of course that is not a statistically relevant comparison, but then again, give me another reason: x86 v. non-x86 emphasis? perhaps, but even that does not seem to be spelled out in HP and IBM's relative strength...all I am asking is for someone to give me some data that dispels my basic argument over the summer: Solaris is killing Sun....
openStorage is fine, go 4 it, make some headway in the growing storage market, and make some efforts to justify the enormous investment that Sun has made in recent years, with nearly nothing to show for it, through open source noise, but don't give me another d*mn openSolaris-influenced press release...enough is enough, you guys (Sun executive management) are holding nice jobs today, and your inability or, more accurately, your lack of will to check your totally out-of-control and myopic CEO is going to result in your unemployment soon...
and all those other Sun employees that have been 10% RIF'ed-like scared in to making any protest to the dominance of Solaris even in the face of all market conditions, forces, and facts, i feel for ya, i have been there, and when i protested something, i lost about 90% of annual income, so i know there is nothing to do...
these blog posts that i make are not intended to be deflating or de-motivating for the truly peerless talent that Sun has in middle management and in the development ranks, but rather, they are intended to be an outlet, for change, for rallying forces to get the CEO to move beyond his own head, and start making decisions that are reflected by factors that are defensible in the marketplace, and stop making excuses, like the pot of gold is just out of reach, and may be attainable in the 'next' quarter...
openStorage is fine; openSolaris is even fine, but it is not the primary mission of Sun, it is a legacy after-thought, though you wouldn't get that impression from the emphasis that has been placed on it, by this losing shoot-the-moon strategy to catch Red Hat...sooner, rather than later, that long-held and morally (i mean, business-wise) non-supportable 10% marketshare of the overall server market is going to evaporate, and instead of Sun commanding a leadership position, it will be gone, like poof, gone...
someone needs to intervene, and i am just in a position to provide the context, the changes, the issues, the arguments, and something of a plan that will be needed to overcome the damage done...someone needs to intervene...
Xeon servers
This is a good announcement for Sun:
http://www.sun.com/aboutsun/pr/2008-08/sunflash.20080819.1.xmland:
http://www.theregister.co.uk/2008/08/19/sun_idf_boxen/Just get that momentum going for some real change, and make the deal that will place the T2 as the compliment to these systems...
stick to standards, bury the hatchet, and start executing, Sun...
openSolaris
Up to this point, I have focused my arguments on the irrelevance of Sun's effort to make Solaris as viable as Linux on the data point of adoption, and so it is...but there is more to the story, and if i can stay awake, i would like to make an effort to explain the opportunity costs argument as to why openSolaris is a problem that needs to be corrected if Sun is to stay in business...
First, the adoption story, which can be rather easily distilled down to numbers, though i don't have the energy to go look it up, it is well-documented that customer, developer, and ISV adoption of Linux is orders of magnitude bigger than Solaris is tracking, and this trend is not going to change, as the recent GPL legal cases only strengthens the viability of OSS, which benefits Linux more than openSolaris...
But it is more of a contextual argument as to why the emphasis on competing at the OS level, with or without OSS, is further diluting the goodwill that Sun had built up, the credibility that it achieved over the first 20 years of its existence, its ability to accommodate evolutions in the corporate IT landscape...that argument, though under-quantifiable, is that there is an incredible opportunity cost of marketing an OS that no one cares about other than Sun managers and customers that do not have the resources to invest in retraining their staff on Linux...
Linux is a global standard, openSolaris is a new project that, unless someone has something to add that i have not thought about, does not have a target user base other than die-hard, Solaris-only developers, and perhaps some customers to try in testing environments that want to utilize multi-core hardware from Sun...is there any way that openSolaris will catch Linux in adoption? No...
Will it have more features that make it competitively advantaged vis-a-vis Linux? Perhaps, as has been argued by Sun's Solaris team with DTrace and ZFS being referenceable as to why developers and customers should choose openSolaris over Linux...does the market appear likely to shift according to demand for these features? apparently not...
So, it is with anticipation that we wait for openSolaris output and Solaris 11's release to see if there are any other features that will re-set the marketplace...i am not holding my breath, and actually feel that this initiative to make openSolaris and the resulting Solaris 11 code more feature rich an investment in badly needed resources that are more needed to address the rest of the Sun portfolio, while customers bet on the far less risky proposition of whether Linux will be supportable by additional vendors, magnitudes more developers, and deployable and integrated with thousands more ISV apps...
I usually, at this point in the thought-process, think of Glassfish, though i understand that openSolaris and Glassfish are not necessarily mutually exclusive, in that you can do both and still operate as a global supplier of IT systems...in other words, the Glassfish team does not necessarily need the openSolaris developers and managers, but the marketing emphasis would certainly be more effective competing in a market where Glassfish has a chance to replace JBoss, WebLogic, and WebSphere in the middleware market, though I really don't think anyone could argue that openSolaris is going to replace Linux...
Where i think the openSolaris development team and managers would be better utilized is making the hardware that Sun sells at a competitive advantage on the T2 chips for multi-core systems, as well as the associated openStorage offerings compatible with Red Hat's Linux...i don't know where we stand in the discussion on whether DTrace and ZFS can be ported to Fedora, but i think the party-line remains that it is not technically feasible; is this a problem? maybe for the customers that have standardized their deployment platforms on these technologies, but not for the vastly larger marketplace that has not bet exclusively on Solaris...
That is the opportunity cost, exclusive of the finite resources that Sun has, including the reality that Sun's resource base continues to decline in the face of less demand for its hardware; and its a virtual cycle, where less demand leads to less resources, which leads to less capability to invest in further enhancements to products, which leads to reduced demand...that is part of the reason why i suggested that Sun is in a
death spiral a few posts back; but it really is more about the myopic focus of Jonathan, Fowler, Green, Anil, and Solaris managers to think that competing at the OS level is priority one, as other initiatives languish, and more importantly, the hardware becomes non-viable because it is not supported with Red Hat....
I just don't see what the execs at Sun have against Red Hat, is there bad blood? if that is even part of the reason, those execs should be replaced, because it is fundamentally killing opportunity for expanding the market opportunity for T2...if there is another reason, like fear that Red Hat will get in the hardware business, i would have to chuckle, that just does not seem to be a priority to Red Hat, i just don't see them competing with Dell, HP, IBM, and other suppliers of their software to customers...what could it be? i just don't get it, and i think Sun owes the marketplace an explanation of why after 5 years of promoting a Solaris-only hardware business even after claiming its support for Linux, there is nothing to show for it, other than some absolutely pointless marketing banter about Ubuntu...
In conclusion, openSolaris is the primary and border-line only reason that Sun is hurting right now...it is diverting resources and attention away from its innovations, leadership, and advantages as a tier-1 supplier...in my weaker moments, i question Jonathan's motivations about driving the value of Sun's acquisition price by a hedge fund to an accomodatable level; in my more realistic moments, i unfortunately come up with an even more inflammatory reflection that he is simply not up to the job, and that he is fundamentally mis-reading the market...what will it take to change course? not much:
a. put openSolaris out to graze as an option for large Solaris shops to maintain support for their older Sun deployments, and offer it as a legacy part of the business...
b. sign the deal with Red Hat, and get working on T2...
c. start the apparently painful but necessary break-up from Solaris-only hardware, and focus on middleware deployments again, as was done in the late '90s with WebLogic, only this time, Sun has Glassfish to maintain account control...
it really is that simple, barring some claims that Solaris managers could come back with on the limitation of working with Red Hat...but the "more features" in Solaris argument is not convincing, the marketplace has spoken, and it would take something much more than ZFS and DTrace-like innovations to fundamentally shift the marketplace in Solaris' favor over Linux...i have looked through this issue, and would like to hear back from someone in the Solaris camp as to why i am wrong; i think this post summarizes the position of Sun Microsystems, circa 2008, and without a change in tactics,
there will not be a Sun Microsystems, circa 2010...
Java marketing
This is the area that I know best, and apologies to my new readers via
JavaWorld, which has invited me to be a "syndicated blogger", whatever that ultimately means, i know that i will reach at least a few more developers via this arrangement to be co-listed on Blogger and
JW...so, when u see that i have posted a new entry, please understand that i have never in my life written a single line of Java code, that i am aware of...the closest thing I come to the world of the
JavaWorld readership is playing with Marimba's Castanet in the '90's, and installing
Glassfish on my laptop...apologies, all around, but thanks to the editorial team of
JW for the invite, and hopefully I will bring something to the table...
Some relevant moments from the recent week have been:
a. Release of
WebLogic 10g Release 3
b. Discussion on
JEE6c. Sun announces availability of
Mobile Enterprise PlatformThere was a time, not so long ago, when these 3 items may have been all the same concept, just different variations on the same product-line, as my time as the product marketing manager on Sun's application server was during the glory-days of
WebLogic within Sun's sales force, and also when
WebLogic was the banner-carrier for the non-product team of Enterprise Java, who worked within Sun, and ultimately the interests of a great many pseudo-decision makers within Sun software, who were pining for a BEA acquisition (or was it vice-
versa, i can't keep it straight, after all these years)...
In the aftermath of the dot-bomb implosion, Sun had to do some soul-searching on the fly, while still performing in the marketplace, and this caused people to actually work, imagine that, the easy days were over, and the top so-called Java people either sucked it up and made contributions to the evolution that would become
Glassfish, or they left, and for the most part, took jobs at BEA, itself...i say this with some reservation, no not really actually, i will never be able to demonstrate that it was actually a cause supported by ex-Sun employees, barring a release of some e-mail thread, but there is little doubt in my mind that some key people checked out of the Sun application server program in 2002-04, and plotted for a BEA merger...
As early as my first days at Sun in '99, I was predicting instead that Oracle was the natural home for
WebLogic, and this must have been clear to others, outside of the opportunists that thought BEA could take-advantage of a languishing stock-price to pick-up a hardware division...so, it is with a bit of relief, inevitability, and satisfaction, that i point to the three items listed above, as the fight to fend off a ridiculous coup negotiated, or at least implied within the ranks of BEA and Sun's Java division, has ended:
Glassfish is here to stay, and Oracle has a viable application server program, to the benefit to all involved, except perhaps a few former employees of BEA and Sun...
good try, guys...
Moving on to business at-hand, the announcements above demonstrate that this is a critical time for Enterprise Java, as competitors and naysayers (one-in-the-same?), continue to deride the "legacy"
JEE platform, and we have little to show for support in the ranks of the purported supporters of Sun's
JEE effort, namely: Oracle, IBM, and
JBoss...all have varying degrees of support for the specification, and all have hedged their bets that ultimately, or at least perhaps hopefully, there will be alternatives to Sun's dominance of the
middleware standards-setting process...I am on the record as saying that
WebLogic is a good money-maker for Oracle, and is finally home, that IBM needs to re-write the entire
WebSphere branded product-line with Geronimo as its base, and that
JBoss needs to get a viable application server back in the market...
Taking a bigger-picture look, it would seem that all of these items will be resolved with
JEE6, as long as these 3 competitors to
Glassfish sign-up, support, and eventually release a compliant
JEE6 product: not a guarantee, however, that this will all happen...Oracle will have its hands full with Fusion, IBM will have a great deal of incentive from Global Services to maintain proprietary lock-in, and i have no idea what
JBoss is thinking or doing these days...so, even though there is a great deal of money in the Enterprise Java marketplace, there are vendors, developers, and even customers that are looking for an alternative, whether that is Spring,
SCA, or some combination of forces...
What can Sun do? I am not sure, though will develop some thoughts on this throughout the next posts that will lead-up to a release of the first draft of
JEE6's final release of the specification, which is rumored to be here by the end of the calendar year...so, that will be my focus for the upcoming months, as well as keeping a close eye on the management and performance of Sun itself, as it needs to stay in business to fulfill its grand strategy around Enterprise Java, as control and stewardship is still impossible under any other scenario, at least as i can see it today...
so, even though i will leave the analysis of particular technologies within the
JEE6 specification to other experts, i presume some of you will find some useful contributions in this blog for how to evaluate whether your skills, your time, your attention, and your aspirations are being appropriately well-represented within Enterprise Java...that is all i can do, and i am open to your suggestions for how to fulfill this objective...
Sun v. Oracle
Note to Jonathan: It's not about data, its about applications. I re-read this blog entry ab/ Sun's self-congratulatory analysis of MySQL's market opportunity:
http://blogs.sun.com/jonathan/entry/mysql_wins_at_linkedinand could only think of how little this guy understands about what OSS is all about, even as he portrays the company that he runs as the pre-eminent open source provider, which it very well may be.
It is with desperation these days that Jonathan talks about market opportunities, and in reality, he is missing the target that needs to be addressed if Sun is to regain account control. That target is Oracle.
The problem with banking on people or customers to take openSolaris seriously is that it ignores the fact that its most acute competitive threat comes from within: in the form of commodity chips from
Intel, running Linux, and Oracle software. Oracle has not been shy, discreet, or even concerned about the nature of their relationship with Sun: it is simply about installed base, as they have no intention
of doing expansionary deals on Sun boxes, no new accounts on multi-core, and no shared revenue on software. I might have been a bit harsh on Jonathan's competitive analysis friday on the earnings call,
with his reference to IBM hardly competing with Sun, but it could be argued that he may be on to something that Oracle is truly the number one threat to Sun's viability. It is time to compete head-on.
The markers are WebLogic and the Database, of course, with Glassfish and MySQL being uiquely positioned to under-cut the hell out of the top-line and margins that Oracle commands on their middleware
and mainstay database businesses. The question is not whether some Web 2.0 company, that has absolutely zero revenue plan, liked LinkedIn, invests in free software with some support contracts, the question
now is how to transfer account control over software pricing and functionality. This is the way to bring many hundreds of additional T2 multi-core customers in to the center of Sun's installed base. Of course, this
assumes that Jonathan ditches openSolaris. It just pisses me off that a simple change in direction, a multi-tenet partnership with Red Hat, and a new marketing chief could turn things around in one quarter. But
I am going to stay on message, and try not to pain myself with the thought that all that i say here could be just window-dressing if the operating systems disaster is not taken care of...
Back to my original note to Jonathan at the beginning of this entry: MySQL can perhaps manage data updates from members who change their favorite colors on their profile, but can it handle asynchronous, multi-
integration points, EJB fail-over application updates? That is the question you have to ask your team, and find out what direction to take MySQL, because leaving it up to the community is exactly what you do
not have the luxury of doing following $1B; leave that to Drizzle. Utilize the core MySQL kernel to bring about a new Enterprise Java focused business plan, and hit Oracle where it hurts, on the middleware
market where standards matter, and where portability is possible. The same could be said about SQL-to-SQL conversions for Oracle databases, but the low-hanging fruit is the middleware market, and then
convert the database customers.
I can only assume that this blog entry will be seen as a positive step for Sun, and not a hit against Oracle, as they are doing fine without Sun. Now, if a real competitive front were to be established, it would be a
delicate dance between the former sibling-like allies, especially with the sales force. But we are talking about the ongoing lifespan of Sun, and sooner rather than later, a tough decision will have to be made in
Santa Clara, and I suggest that getting Oracle off the price list is just the sort of move that makes things more risky, but certainly more interesting, and with one move, a potentially game-changing move toward
relevance again. I have said it about a dozen times before, about this price-list issue, but it is only a matter of time, until Oracle Linux takes away all Sun accounts; perhaps, maybe I am just dreaming here, but perhaps
it is time to fight back. I honestly don't know what to say anymore, I know the problems are plentiful, and complex, and long-term, but I have run from such problems before and they only get worse over time.
Wishful thinking on openSolaris, partnerships with enemies, and bad leadership are no excuse. It is time to change, and I am generally pretty inexpensive to bring-in to correct the middleware problems. I mean
what else is there to do: bring in McKinsey to advise on where to do the most effective R.I.F.'s; c'mon, do better than that...
JEE 6
Well, lets just say i do have the capacity to move past financial news and make some comments on something that is going right at Sun: of course I am talking about the impending release of Java Platform, Enterprise Edition 6 (JEE6), as this seems to be the only thing that is consistently executing in a leadership position right now, and that includes the competition with Spring and .Net...there is not much in the way of comparing JEE6 with Spring 2.5/3x and .Net 2/3x, as they are each fulfilling their own niche, so i won't do that justice, i'll leave it to a discussion on TSS to vet that one out, but I will talk about what Enterprise Java means to Sun's overall business, as I have made some references to it in previous posts...
For your background reference, details about the JEE6 platform can be found here:
http://jcp.org/en/jsr/detail?id=316There are a lot of big questions remaining over what this release will actually deliver on: with JEE3, there were some JAX web services integration via the Web Services Pack; JEE4 formalized that JAX integration; and JEE5 re-wrote the EJB and JPA specs to make it more viable to more developers...now, with a focus on making JEE6 more palatable to more developers, especially at the "web-tier", there is this thing called profiles that is being introduced:
i think this is a mistake, but i have been wrong before, and do not feel like fighting it, in the limited scope of what such a protest would make...essentially, profiles means that after years of arguing with my former colleagues at Sun about calling the Sun Web Server a 'light-weight' application server because it served up JSPs and Servlets, that is exactly what the JEE6 expert committee is about to certify, with the real significance being that Tomcat will be able to call itself "JEE-compatible", as will the Spring Application Platform...
why is this important? well, to state the obvious, it destroys the very value proposition that made JEE the de-facto Internet application development paradigm, as customers faced a far-less costly portability exercise...for instance, develop a JEE application in Glassfish, with or without EJBs, and there is no explicit guarantee that it will work with other so-called application servers, as it is up to the vendor to determine which 'profile' to support...
why would the expert committee shirk its main responsibility of providing a standard specification that allows for full compatibility with Java? arguments have come from far and wide, and i wouldn't mind hearing them again, as i don't get them, and i even boil it down to a political manoeuvre to splinter the JEE market...is this paranoia? perhaps, but i have yet to hear a valid argument that dispels this possibility...
o.k., so with profiles, you can supposedly write a standard servlet and have it run cross-platform, and perhaps this weird-thing called 'EJB-lite' might apply to this principle as well...but anything that Sun's main customer base is going to do with JEE, in particular banks and telecos, will be utilizing enterprise features such as JPA, JAX, and even full EJB - - what then? it is up to the development team to choose what works best for them, but for the platform architects, a much more informed analysis is needed to know what will be essentially "standard" and what will be a profile...why is this good? as stated, the argument is that more developers (meaning more novice developers) will be able to utilize the power of JEE...
as u may have gathered, i am less than convinced that this will result in better options, in fact it may force developers to move to a more standard specification in the form of Spring, as at least they will know what is coming, what is portable, and what platforms support it...and this really comes back to Sun...what on planet earth is going on at the company?
I have been pretty clear about my lack of support for openSolaris as a driver of Sun's business; in fact, it is JEE that is the driver of Sun's business, as Enterprise Java deployments are custom-made for the T2 hardware business, and have additionally been the mainstay of Sun's revenue for the past decade...why splinter this effort? I know that the JCP is a 'community' process, but there has to be a reason why Sun has resisted completely turning over the direction and final decision-making in favor of having final veto...my thinking is that the veto might be best used to strike down the profile effort, even as it would result in an unprecedented uproar from many of the JCP's most prominent members...
or would it? SpringSource would be upset and would more vigorously disparage EJB and also to JEE, to a certain extent; Oracle and IBM may be upset with Sun's heavy-handedness this late in the stage of development of JEE6, and Apache would of course be up-in-arms (for something other than licensing this time), but really who would be affected by this: so-called developers who are waiting in the wings for profiles? i doubt it...
I am nearly sick of hearing myself think about the lack of management savvy at Sun, and it would seem that the executive team has more pressing things to worry about, such as a diving stock price, and a diminishing installed base, but it would signal to the Java marketplace, that compatibility, portability, and ultimately, re-use are the driving forces behind JEE; with profiles, this argument may die...i am a bit on the tip of being negative about everything coming out of Santa Clara right now, and i am not convinced that JEE6 is the area where concentration should be directed...
but i do know that JEE6 is important, Glassfish is vital, and the Java developer community is the only constituency sticking it out with Sun, so perhaps some leadership would be welcome...I don't know what else to do about this company, it is so severely entrenched in myopia that i cant imagine what meetings are about except the day-to-day tactical initiative to keep customers from jumping ship, but somewhere, at some point, a decision will have to be made to release JEE6...
considering the inability to get anything else right, maybe this is a good, safe place to start: eliminate profiles and market JEE with Glassfish as the whole center of the business, i can't see something else with more upside, i am open to alternative suggestions...
reflections on friday
I don't even care to apologize for some of my analysis on friday's Sun Microsystems FY08 Q4 and full year earnings call. I guess what I would like to do is provide some background on what has brought me to this place of
seeming uni-focus on one company. There are 2 key explanations for it:
1. I grew up on Sun. Much like my internships in college and graduate school, I was early-on influenced by the magnitude of this seemingly innocuous strategy of Sun's to sponsor a programming language that would at once be viable,
while at the same time being magnanimous in its implementation, basically by allowing it to run on something other than its own Operating System: Solaris. Essentially, it was 1995, and I was struggling to figure out what to do with my life,
as undergraduate days were passing by and I was getting sick of politics, which is the area of study that I had chosen (no regrets). Alongside Netscape, which to any college-age student at the time, was the ultimate in freedom of expression,
as one could now access the world wide web without restrictions, Sun introduced Java and its bastard-incarnation JavaScript (albeit this was actually Netscape's way of making Java perform). I remember working out the early drafts
of a business plan that was undoubtedly inspired, required the use of, and ultimately bound to the concepts of Java's multi-platform, re-use, developer intrigue value proposition. This was spring of 1996, and it was a plan I still have
somewhere.
Essentially, it was to allow students to write applications that would be accessible on-line, and they would be paid for the use of these programs. I am slightly embarassed to admit this, as I know of several incarnations of this same idea
that have been attempted, but it was 1996, and it pre-dated the official release of Marimba. I remember working on this plan, as it included some "cloud-like" attributes, such as an operating system for the applications the students would
write (essentially, what would later be known to me as an application server), a mechanism to deliver the applications in a customized manner (what would later be known to me as Castanet), and a customer-base that would pay for
usage, not just to have something pre-loaded on their computer, as Microsoft had established with all of their profitable products. Though my focus on the 'education' market, of empowering students to write applications was a bit
cumbersome for figuring out how and when students would do this plan, it was my catalyst in to the world of pre-built software components, that would form the hallmark of my days leading up to and ultimately as an employee of Sun,
in the early form of the Sun-Netscape Alliance.
I am extremely proud and fortunate to have had the opportunity to work on Netscape's campus, the original brown-stone-ish 2 and 3 story buildings in Mountain View that I believe are now occupied by that networking company that
began with a "V", I can't recall, but whatever, anyone who is anyone in the Valley, knows the old Netscape campus, and not only did I get to work in building 23, but I got to actually work with Netscape employees, who by then in
September of 1999 were AOL employees. I was hired by a Netscape middle manager, and the basic truism for Netscpae middle-managers, as far as i could tell, is that they were by far, without question, the coolest and smartest
people in software. The browser-based middleware strategy of Web, App, and Directory Servers was the reason why BEA, Oracle, and IBM even knew how to compete with Microsoft for developers, it is because of the immense
talent at Netscape that Java even has a chance today, even as we long-ago have forgotten about that company called Netscape.
My final year of graduate school afforded me the opportunity to travel to the Bay Area a couple of times, first for my cousin's wedding, and then NYE on my way to SE Asia, and then for Spring Break. During these social visits, I
made every effort to talk with people in the Java market, as I knew that if my educational application development start-up idea was not going to initially work, I wanted to work on what I thought was the only place that I could:
in the very early-emerging Internet software market. I talked with Oracle, I talked with Silicon Graphics, I talked with Network Computer (that thin client operation sponsored by Ellison), and others, but I was somehow, through the
distribution of a short paper I wrote on the to-be-announced Enterprise Java platform, able to get some informational interviews with Sun people. I basically made my way through the NetDynamics ranks of Sanjay, Rakesh, Ratnesh,
and the guy who changed my viewpoint or at least gave me the confidence to pursue my career path, Dan Graves.
I still remember sitting in his office, and his drawing up the basic constructs of the Enterprise JavaBeans (EJB) component-model for re-use across multiple platforms. It is relatively unclear to me if he was the only one willing to dream
that big, or was the only one willing to talk about it, but it was in-line with what I had been thinking about for student development, and Dan was suggesting it was the only reason that NetDynamics had just been acquired by Sun. This
was following BEA's acquisition of WebLogic, so the market for application servers clearly had some potential. I figured it was where I needed to be, and after much harassing of Yuan Huntington of Netscape, I got a job in the fall
after graduating from William and Mary. No experience, no programming knowledge, no connections beyond what I had made in 1998-99 throughout the NetDynamics org., and yet Yuan took a flier on me pretty much based on my
excitement for the possibility of a new kind of "operating system" called an application server.
I started and remained at Sun for about 4.5 years, never once changing my title as the Product Marketing Manager on Sun's application server: Netscape, iPlanet, Sun ONE, and finally Sun Java System. I have many little stories to be
filled-out over how I maintained that position on a relatively crucial piece of Sun's future business strategy, but suffice to say, I left on my own terms, under huge amounts of duress, that does not eclipse the fact that today Glassfish lives,
and my apologies to all the Netscape, iPlanet, and Sun people who I worked with, but there would be no viable application server program at Sun without me. Arrogance, myopia, and inertia aside, I gave my career, sacrificed my
well-being, and stood up to injustice in the name of that damn product, so i could give a f*ck today what people at Sun who think I am being inflammatory or harsh on the CEO think. I grew up on Sun, and I think I have earned the
right to talk about it today.
2. Sun is in a death spiral, and it is simple math or at least information technology math, that once you take away the aura of inevitability, such as for customers to start to think: 'hmmm, perhaps Sun won't survive,' the wheels come off,
because deals just melt away. After Friday, that is exactly what is going to happen over the next couple of quarters. It is going to be ugly. And it is because of enormous arrogance, that is probably better understood as stubbornness,
because the constructs of a deal are in place over the JDK between JBoss and Glassfish that would greatly enhance the Enterprise Java portability, and re-use argument, as well as the most obvious point of all: that Sun needs Red Hat
Linux as a lifeline to sell the SPARC "Niagara" T2 hardware. You can reference my posts below to get some further context for this argument, but it bears repeating: Sun will die unless it ditches Solaris-only mindset on their most
valuable hardware.
I hope you can see why the two points are related, and that is why I have begun blogging again, as anyone closely following Sun for 10 years, as I have, knows what is going on. It is time for renewal of the value proposition that said:
'there is no pride in placing the advantages of a community-developed product over the merits of our own internally-developed product.' Solaris and Java are on two completely separate trajectories, and Sun will run out of accounts
to call-on, readily available assets, and ultimately it will lose its installed base, if it continues to be stubborn. I have made some bold characterizations, I don't really care for some of the people feeding the openSolaris mindset, but this
is more than personal, it really comes down to making the right business decision. And that means making it before it is too late. I can't think of anything else to do but this:
- sign Red Hat to Global Partnership across all issues
- make Glassfish the only supportable application server for GSO
I have said it before, but it is worth mentioning one last time, if steps are not taken soon, there will be no niche, no market, no product-line, and no base for Sun to justify corporate goodwill, future growth, and developer sentiment, it is
time to cast aside the stubbornness of the current regime, and save this company...
$1B: why not?
Hey, if you got no business model, other than some b.s. about open source, especially related to your completely irrelevant operating system, but have left-over proceeds from the dot-com run-up, why not buy some shares and make the complete dis-integration in confidence with the company's management team not become a full-scale sell-off of what is left of the artificially supported stock price?
Summary of the call:
a. Q1 is going be a disaster, but we have no reason to give why that is, even as the rest of the industry stabilizes.
b. BRIC and other developing countries are the future, although its about a 90-10 comparison with the U.S. being the larger number (not good, to be betting the company on BRIC)...
c. Product-line is moving away from the openSolaris positioning, and moving to openStorage.
d. Cash balance is on-track to be gone within 2 fiscal years on the current trend, unless you believe that no additional acquisitions are needed.
e. Gross margins are deteriorating, but just trust management, R&D innovation will somehow bring that back...e.g. Thumper needs to trump StorageTek product-line (but don't mind that its currently $120M v. $2B comparison)...
f. Niagara is based on the success of openSolaris (never mind that no one cares about this OS)...
g. Summary of Jonathan's competitive analysis: 'We really don't compete with IBM; they OEM Solaris, so we're more partners than competitors'...my response: bahahahaha...
h. Jonathan is bitchy this morning, must be the early wake-up call...
i. Finally the billion-dollar question: why is Sun missing expectations? Answer from CFO Lehman: we think other tier-1 vendors (not named) are struggling; large U.S. customers are leading to weak numbers; Jonathan: no answer, didn't even answer the freaking question: weak effort from the CEO...
j. Sales execution or Product issue? Installed base not coming through? Bad acquisitions? Answer from Jonathan: growth in BRIC and related, blah, blah, blah...also, Large Financial customers are to blame; U.S. downturn is the main reason why we can't perform..."we just got to grow": thanks, for the update...
k. Main point from CEO and CFO: IT purchases are being tabled, we need to grow outside of 'mature' economies; my take: apparently only Sun thinks there is a recession...
l. CFO is way bitchy...
m. (I don't sit on many other companies' calls, but this is turning nasty)...
n. Jonathan: "Academic environments for MySQL"; my take: give me an f'in break, your a F500 company...
o. Main base for revenue is U.S., which is expected to be down until 2009, so Jonathan's model for building the base back: partners...my take: all the best, there...
p. Commentary from Doug: no mention of Red Hat partnership = irrelevance for Sun...
q. "4 pronged attack": what the hell is this? i assume this is the 4 S's: Servers, Software, Storage, and Services, but still continual fluff from CEO on "value proposition"...
r. Commentary from Doug: Jonathan is not prepared to lead this company at this stage in its life, it needs someone less 'visionary' and someone with more turn-around experience...
s. Blah, blah, blah about Solaris unstated revenue on CMT and Thumper sales; no one cares anymore about this analysis...
t. my line went dead on the lack of hardware pull-through with MySQL deployments, essentially a $1B acquisition is not bringing in anything on Sun hardware; my answer to this: shocker...
u. end of call...
Analysis: see letter r. above; end of story...
