Skip to content

What Would SOG Do?: Views on Sun

Following Sun’s announcement last quarter, the tenor of the inquiries we received about the vendor changed, and changed dramatically. Post-earnings, and subsequent to the announcement of a 18% reduction in headcount, third parties and media types alike shifted their focus from the firm’s products to the firm. What was to become of Sun?

The firm is no stranger to questions about its future, of course. As one of the beneficiaries of the dot com boom, it was punished disproportionately for that implosion. Until that point, the organization was a high flyer that could do little wrong, with its operating system the enterprise standard and its workstations – the basis for its old ticker symbol, SUNW – an iconic piece of technology.

That the road from there to here has been rocky goes without saying. The question now is simple: where does the organization go from here? This entry, like Tim’s that inspired it, is one attempt to answer that question. Given that I am not privy to the machinations of either Sun’s board or its large institutional investors – in whose hands the fate of the firm rests – the following should be considered an editorial, at best.

As is my wont, I’ll conduct this discussion in my typical fashion, as a Q&A. If there are questions that you think I’ve missed, feel free to submit them afterwards, though I may or may not be able to answer them.

Q: Before we begin, do you have anything to disclose?
A: Certainly. Sun is not only a RedMonk customer, they are one of our longest tenured clients, as well as one of the most committed – historically, anyway – to the RedMonk cause. I also know and respect many of the firm’s employees, from Jonathan Schwartz on down.

Q: Would you like to preface your remarks at all?
A: Indeed. There are several caveats to be aware of, among them:

  1. Context:
    While the following is obviously a commentary inspired by a series of poor showings from the firm, financially, Sun still has a few billion in the bank, and to take their run rate down, they’re losing almost one in five people. It’s also presumably the case they won’t be paying for MySQL and StorageTek every quarter. The following commentary, then, should be taken in the spirit it’s intended: as constructive suggestions for those mapping Sun’s future – not a sky-is-falling obituary.
  2. Cuts:
    I, like many others commenting on the subject, will recommend that Sun decommit from certain assets. This is, obviously, easier said than done. First, because some of the businesses are generating some cash and involve customer commitments. But also because some of the recommended activities – transitioning OO.o to a foundation, for example – are actually more expensive in the short term.

    You should read the below then as a set of general guidelines, rather than an “everything must go, and go now” mandate.

  3. Numbers:
    I don’t have them all, obviously. So it’s difficult for me to say with precision what Sun should and should not do with its various assets. That said, the following are what I believe to be reasonable assertions based on the available evidence and my experience covering the firm.
  4. Products:
    I’m generally in agreement with my colleague when he says that Sun’s product catalog is as strong as I’ve seen it. It probably sounds weird to hear that about a firm that is currently being bludgeoned by the market, but take a look around. MySQL is the volume database, the x86 line is a growing business, storage has been reinvented first with Thumper and now with Amber Road, and Solaris – still – is technically differentiated. Having good products is not Sun’s problem.
  5. Share Price:
    While many of the analyses I’ve seen of Sun recently focus on its share price – understandably, given its current levels – I believe that to be an extremely imperfect metric for the performance of a firm. As defense of that claim, I recommend checking out MSFT‘s 5 year share price fluctuation: that, for a firm that is, effectively, a mint.

Q: On a macro level, what are the potential outcomes for the firm?
A: As with any entity in a similar position, there are multiple possibilities. The most likely, in no particular order, are these. First, they remain a viable, ongoing concern following the planned streamlining or rightsizing of their business. Second, they are acquired for their assets. Third, they are acquired by an entity strictly for the purposes of selling their assets. Fourth, they burn through their remaining cash and are no longer viable.

Q: Which of those possibilities is most likely? Can you provide odds?
A: That, in many respects, is the $64,000 question. Unfortunately, I’m in no position to collect the cash: I cannot predict with any certainty what the future holds for Sun. Not strictly because I abhor the business of prediction generally, but because the answer depends on data that I do not have at this time. We know they are cutting 18% of the workforce, for example, but I’m not privy to the details of where the cuts are coming from, and what if any products are being cut. Ergo, I am unable to speculate with any degree of certainty how the firm will perform going forward.

Q: How did Sun get here?
A: That’s a complicated question, one that cannot be answered simply, or in the context of an entry like this. I will say, however, that some of the difficulty stems from Sun’s culture, which is engineering driven.

Q: Can you explain?
A: Certainly. Engineers, as the cliche goes, are driven to innovate first, and seek markets second…if at all. This behavior, in many respects, has come to characterize Sun. Like the fabled Xerox Parc – though, fortunately, with more business savvy and commercial success – Sun’s culture is engineering dominated. Innovation is valued, to the extent that under McNealy’s watch, necessary headcount reductions were postponed in search of game changing innovations. This culture stands in stark contrast to other businesses in the industry.

IBM, for example, is in many respects the polar opposite of Sun. While it is no slouch in the area of engineering and innovation, the firm is intensely business model oriented – at the expense of engineering, if necessary. This approach has its costs: by prioritizing immediate profits, subtle or more difficult to perceive markets are either ignored or not properly leveraged. Think operating systems and web search, two markets that IBM could have dominated, but failed to demonstrate the opportunity for immediate profit.

Sun, on the other hand, has long funded innovation for innovation’s sake. When it appeared to everyone else – myself included – that further investments in Solaris were a waste of time and energy, the firm doubled down. And was rewarded, from this point of view, with a highly differentiated, if still regrettably inaccessible (though that condition has abated), product. Similar investments in tooling, on the other hand, have yielded innovation but little obvious opportunity. As we’ll get to later. Their valuation of their own technology, as well, has led to an occasional NIH complex, which poses its own challenges.

Q: So you believe that Sun should become more like IBM?
A: Based strictly on the financial performance of the two organizations, it’s fairly clear that Sun could learn something from IBM, yes. But frankly, so too could IBM learn from Sun. The point here is not to advocate one approach at the expense of another, but to recognize that both have their time and place, and that an organization that is dominated by either an engineering or business model culture is likely to be one sided. Ideally, both cultures are leveraged and applied in a balanced approach.

That’s the goal, anyway.

Q: Do you subscribe the conventional wisdom that assigns the blame – either all of it, or part – to Sun’s CEO, Jonathan Schwartz?
A: I do not. The first time I met Jonathan was in a Boston hotel room along with some of my ex-Illuminata colleagues, as he took the reins of the software organization. He impressed me then with his intelligence, insight and pragmatism, and I’m not of the opinion that his intelligence has declined since that time. I have not agreed with some – perhaps many – of his decisions, but I’m also cognizant of the fact that the firm’s portfolio at present – dire though the firm’s financial condition may be – is better than it was under McNealy’s tenure. Significantly so.

Does he bear responsibility for the firm’s decline? Of course. As any chief executive would, under similar circumstances. But I’m not sure who could have done better with what he was handed.

As an example, around the time Jonathan was ascending the ladder of Sun’s software business, James and I did an assessment of the software porfolio for Sun, and frankly, it was horrific. There were literally dozens of products strewn across the canvas of the organization, few of which had any consistency in terms of delivery, business model, compatibility, target platform, or otherwise. The story today is improved, if still highly imperfect.

The other question I think people don’t ask when they blindly advocate a replacement of the CEO: who’s your replacement?

Q: So what would you do if you were in Jonathan’s position tomorrow?
A: Everything for me would start with numbers. These numbers, specifically. The second slide, in particular, would dictate much of my approach.

As Cote has said, there are other businesses such as Adobe or Microsoft that can afford to subsidize significant, non-earning investments with franchises that print money. Sun does not, however, happen to be one of those businesses today, if ever it was one. Businesses that don’t make money or show at least the prospect of earning money, therefore, should be either cut or dramatically scaled back.

Regardless of how impressive the technology is.

Q: Can you provide examples?
A: Well, much of the impact would depend on variables that I do not have – specifically the headcounts of the various products, their expense structures and so on – but two that would seem to be luxuries at this point, to me, would be NetBeans and OpenOffice. Neither, you’ll note, appears on the Line Items spreadsheets linked to above, meaning they do not represent a potential source of income, meaning that they should be considered – in my view – a luxury. And Sun is not in a financial position, at present, where it can afford luxuries.

Q: What about the arguments in favor of those products?
A: Well, let’s take them in order.

So, NetBeans. It’s no secret – either within Sun, or to those that have spoken to me on the subject – that I’ve never understood Sun’s commitment to NetBeans.

In the early days it was, frankly, because the product was poor. But even as the offering improved to the state we find it today – highly competitive, even differentiated in certain markets – I remain convinced it’s a poor investment. Which is one area where Tim Bray and I apparently do not agree. Tim argues that Sun should focus on the a Web Suite:

Therefore, Sun should adopt a laser focus on building a Sun Web Suite and becoming the Web application deployment platform of choice. It’s a large space, a growing space, and one where we can win.

While I could quibble aspects of that claim, I think he’s largely correct: that is an opportunity, and one where Sun has interesting pieces. What I remain profoundly unconvinced of, however, is that tooling remains a critical portion of that opportunity. Where critical reads as “worth investing increasingly scarce resources in.”

Let’s accept as a given that NetBeans is profoundly improved from the Swing-slow toolset of a few years back. Let’s further grant that Sun has been market responsive of late, and gifted the toolset with some excellent dynamic language capabilities. My question is this: how does this make Sun money?

Is the attach rate for Sun infrastructure – the stuff that people actually pay for – that much better for developers that use NetBeans versus, say, Emacs? Because that is, in many respects, the competition. I have seen little evidence that – outside of Microsoft’s hermetically sealed environment – a tooling story is intrinsically complementary to an infrastructure story. Most of the web developers I know use text editors (BBEdit, emacs, TextMate, vi, etc), not IDEs. And those that do use IDEs tend to use Eclipse, not because it’s necessarily better, but – like Firefox – because of the ecosystem around the product.

No matter what they use, however, the tooling has little if any direct impact on the infrastructure that backs it. Many of the NetBeans users I know are writing to a LAMP stack running on Dell hardware, while back at Major League Baseball, the staff was using Eclipse to write to a Sun portfolio. And if the tooling is not a lead for revenue generating infrastructure commitments, where, precisely, is the return on Sun’s investment in the platform?

If you’re still unconvinced, consider that the dominant cloud providers of the present – Amazon, Google App Engine, and – have no tooling story of their own, and yet are successful. Why then would – or should – NetBeans be predictive of success for Sun from an infrastructure perspective? My opinion is, obviously, that it shouldn’t, and will not be.

If NetBeans isn’t making money itself, and isn’t cementing sales elsewhere, what is its purpose?

Q: And OpenOffice?
A: is nearer and dearer to my heart, as I’m still a user of the product – though much of my office productivity time these days is spent in Google Docs – and 3.0 is a remarkable achievement, performance-wise. As I’ve documented elsewhere, 3.0 will cold-start open a document for me in just under 3 seconds; contrast that with IBM’s Symphony, built on an older version of the OO.o codebase, that takes just under 13 seconds on the same document.

But while Sun has fought the good fight with OpenOffice, and made possible a raft of new innovations by combining with IBM to champion the OO.o derived ODF format, the time has come to let the effort sink or swim on its own. Over three years ago, I called for for OO.o to be transitioned to a foundation, and was more or less dismissed. Here we are, in 2008, and the product is still funded – although at reduced levels, it must be said – and still not making money: note its absence from the Line Items spreadsheet.

What then is the benefit of the product? Convincing people that Sun is serious about open source? I think most people have gotten that memo. Visibility? Possible, if you think that people pay attention to splash screens.

Here’s Jonathan on the subject of OpenOffice:

As for other high value distribution assets at Sun? I just read one analyst report questioning whether anyone actually used OpenOffice. We happen to run Sun Microsystems on OpenOffice – more importantly, it’s used across the world, and we’re now commercially licensing it to brand name companies wanting to save big dollars on office productivity.

To put some data around its popularity, last week, we distributed more than 3,000,000 copies of OpenOffice 3. Downloads are accelerating, giving us a reachable user base we estimate to be between 150,000,000 and 200,000,000 users – a global recession will amplify OpenOffice adoption. And 100’s of millions of users drive a lot of foot traffic. An auction’s afoot (no pun intended) to see who we’ll be partnering with us to integrate their businesses and brands into our binary product distribution – the possibilities are limitless: people tend to print those documents, fax them, copy them, project them.

That analyst report, it should be noted, was not ours, because the one thing I would not question would be the adoption of OO.o. It is, frankly, everywhere. But unlike, say, an enterprise database (e.g. MySQL), the odds of most people paying for an office productivity product are slim and none. Meaning that the return on Sun’s investment (whatever that may be) in the product is, at best, problematic.

Like NetBeans, strikes me as excellent technology, but excellent technology without an immediate revenue opportunity for Sun. Meaning that it has to be cut.

Q: What else would you cut?
A: Like Tim, I am not a believer in the client-side Java technologies. JavaFX always struck me as an also ran in that market, particularly once I heard the story of its ascension. While Flash and Silverlight are guaranteed places at the table by virtue of their respective positions, Java simply isn’t a major factor in the rich client landscape.

Which may not even be that much of an issue, because – again like Tim – I am not a believer that the world is going to be all that rich. I’m an early adopter, and the only RIA that I regularly use is Twhirl.

So are continued investments in JavaFX worth it? Not in my book.

Q: What is Sun? A hardware company, a software company, or both?
A: Hardware would probably be the most defensible argument. For FY08, the Systems folks reported 6.5B in billings, the Services crowd 5.3B, the Storage guys 2.4B, with the software kids bringing up the rear at 644M. Assuming that I read the line items correctly.

Q: So should Sun spin out its software business and become a hardware company?
A: That’s for others to say, but that doesn’t make much sense to me. It’s easier – by far – for Sun to differentiate itself on the software side than the hardware side. Not to mention that hardware businesses are declining, generally. Meaning that the barriers to entry, even if they haven’t perfected the model yet, are greater for their software business than their hardware business – just ask the BTRFS and the SystemTap people.

The immediate concern, of course, must be revenue, and revenue generation. But spinning off a differentiated asset for a potential short term profit to preserve a difficult to differentiate hardware business represents a logic that’s hard for me to follow.

As product lines like Amber Road demonstrate, the firm is at its best when it’s leveraging the respective strengths of the different lines to innovate. Spinning one out would seem to preclude that as a straightforward approach.

Q: Speaking of combinations, what about the cloud? Do you believe that Sun can and should play there?
A: Can they? That remains to be seen. As to whether or not they should, I differ again with Tim, who said “I’m not convinced that Sun can succeed as a large-scale supplier of cloud services, and I’m not even convinced that we need to,” as well as with other Sun executives who’ve told me in the past that they didn’t intend to be “in the hosting business.”

True, the cloud margins are slim. Perilously slim, perhaps, for a company as market impacted as Sun happens to be.

But as I argued in February, I do not believe that Sun can afford to not be a player in the cloud market:

Will said future, for example, include a computing landscape largely dominated by four or five big computers – Amazon, Google, Microsoft, and so on? And if it does, will Sun be one of those big computers, or merely a supplier?

Narrowly focused efforts like aside, Sun has to date shown little ambition to be the former. Which would worry me, because success with the latter assumes to some extent a reversal of the general preference among the big computer folks for non-premium, whitebox hardware. If there are going to be four or five big computers, and you don’t have a history of selling into them at volume absent significant subsidies, I’d consider that an issue.

As I’ve been saying publicly and privately for some time, I think it’s important – vitally so, in fact – that the vendor that coined the “network is the computer” tagline offer a network of its own.

Put more simply, let’s consider two markets: the enterprise and the web/developer.

The former are buying from Sun, still, but at a declining rate: hence the challenging results last quarter. Adding insult to injury, it’s possible that – in a vicious cycle – the enterprise may become a more challenging market given those results.

As for the web/developer market, they have not, historically, been Sun customers, and while they’ve purchased Dells in the past, the trendline here is increasingly skewed towards cloud platforms such as Amazon’s.

So where’s the entrypoint for Sun to make money? MySQL, perhaps. OpenSolaris, if the technical differentiation is properly leveraged and the comfort level factor can be overcome. Glassfish, possibly. But just a minute ago we discussed that – from a revenue perspective – Sun was not a software company, but one driven by hardware.

That tells me that it cannot afford to cede the hardware market to the likes of Amazon, because few of those providers have demonstrated an affinity for premium hardware. Google and its brethren are content to assemble their own gear from whitebox hardware providers, leaving Sun to service the less capable providers. Being an arms provider in a market content to manufacture their own weapons is not a future that I would envy.

Meaning that I would invest – and invest heavily – in the cloud.

Q: What about the comments about the field sales force? Do you think Sun needs a change here?
A: Sun’s field sales force is, as noted, equipped to handle a very particular market: large enterprises. This they do reasonably well, or as well as can be done in a market like this. Having met with Peter Ryan last week, I feel reasonably confident in saying that Sales are in good hands, and that what can be done will be.

The reality, however, is that this market will be challenging for the next several quarters, and there’s a reasonable chance that it becomes significantly more so. There’s only so much a sales force can do, at this point: the market is what the market is.

Of interest will be how Sun does or does not adapt its sales process to sell to non-traditional customers. While it’s long been a goal of Jonathan’s to enable Sun to sell to customers it never touched directly – a goal that I applaud, incidentally – its execution on this aim has been, frankly, less than impressive. Startup Essentials has made progress in this area – they stopped calling me, at least – and programs like Try and Buy are very creative, but collectively it’s not enough. When I speak with startups today, Sun is not top of mind.

Amazon, on the other hand, is. Without a massive direct sales force. From that I conclude that you should leave the sales force to its prescribed job of selling to the big guys, and give the other market what they want: a no-barriers-to-entry cloud product. Maybe with the help of some Jeff Barr-like evangelists.

In other words, the problem with Sun sales is not – to me – a logistical problem, but rather an inability to provide the market what it wants. That’s where I would start.

Q: If all or most of the above is tactical, where would you place your strategic bets from a revenue perspective?
A: My strategic bet with respect to the software business – outside of the traditional markets – remains the same: network services. Maybe if I came up with a cooler name for that, people would pay attention…but I digress.

The fact is that many of Sun’s businesses – like other open source businesses – struggle to convert users to paying customers. Experimentation abounds with hyrbidizing the technology, blending free and proprietary software, but the long term play from my perspective will be value added services that are delivered via the network. You use some bits, and pay for the value added service that is delivered on top of them. Think Akismet.

Longer term, were I Sun, I would be scrambling for opportunities to layer for-pay network services on top of my heavily adopted open source platforms. That’s a logical buy trigger for users, and a potentially lucrative recurring revenue stream for Sun. I’m not going to pay MySQL to run RedMonk’s databases, because the software is well written and does what it is supposed to do. I would pay MySQL, however, to watch my database telemetry, give me reports on performance, spikes, pending problems and issues, and the like.

Q: What would you tell Sun generally?
A: Most obviously, go private if it can be arranged. Getting regularly flogged, publicly, for trends that will take time to revert is spectactularly unhelpful – as I’m sure everyone there is all too aware.

But more importantly, as Sun is an engineering first firm: that the best technology doesn’t always win, and that there’s no room for emotion in a market like this one. Determine a product’s strategic importance and future profitability ruthlessly, and then cut anything that does not impact the bottom line as expeditiously as can be arranged.

This is unpleasant, and will mean political infighting, but it needs to be done.

Q: What do you think Sun will look like, on the other side of the proposed changes, as a firm?
A: I’m less concerned with the overall picture of a firm than I am with the results. Ideally, Sun would be positioned as a leaner entity focused on profitable or potentially profitable businesses: whether those are perceived to be web, enterprise, consumer or whatever is less critical than its ability to perform for the financial markets at this point. Tim spoke in his piece of certain businesses being “distractions,” and there is certainly a risk that an unaligned portfolio of assets means that the big picture story is more difficult to relate. That said, I tend to think that some of that is the byproduct of an unnecessary coupling of the separate product stories. I also think that – just like in baseball – a little winning would cure a lot of ills.

Q: Do you believe the firm can survive?
A: Absolutely. I think it will be challenged to “stay the course” from a product portfolio perspective and succeed, but with the appropriate adjustments I see no reason that the firm can’t recover and emerge stronger for the experience.

Categories: Business Models, Open Source.

Comment Feed

35 Responses

  1. Generally agree with this analysis although I might be inclined to relatively emphasize different points. At this point, things like Open Office funding are pretty small potatoes in the scheme of things–not that I’m going to argue too hard with your conclusions, especially WRT OO. Bigger issue seems far more execution (including around the boring but profitable stuff) than strategy. And it certainly felt like they paid $1b for STK which may or may not have been a good idea, but once they did they largely lost interest in it.

    Gordon HaffNovember 26, 2008 @ 9:25 amReply
  2. Stephen, interesting that you didn’t mention Java directly – that’s still one of the more valuable brand names Sun “owns” and is probably the Sun technology that touches the most people.

    It’s omission seems STRANGE.

  3. I agree with you on NetBeans and OpenOffice. If they’re any good, the open source communities should be able to continue maintaining them and making them better.

  4. @Gordon Haff: thanks, and i agree that one can turn the dial variously on the properties.

    the macro conclusion, however, is one that i think we’d both agree on: Sun needs to aggressively pare back, and kill off a few sacred cows.

  5. Great piece Stephen. We’re in pretty clear agreement.

    Rich- we don’t see it the same as you. If i was Jonathan i would never have changed the stock sticker to the name of a technology that never drove significant sales at Sun, and imho remains somewhat of an engineering albatross around its neck. Things have dramatically improved – but at this point Glassfish is probably a more valuable brand. Its somethng people want and they will come to Sun to purchase subscriptions for it.

    A valuable brand? Take another look at Sun’s software revenues. And no I don’t buy the pull through story. Sun is not currently kicking ass and taking names in the media storage business because of Java- its because the storage products, with ZFS and low cost server infrastructure, is a mighty compelling offering.

    Even the wireless telco market, one of the key Java spaces (my N95 Google Docs app is written in Java) – isn’t creating significant Java related revenues. We pay Google for the docs, and i pay Vodafone for the service.. and yes i know Vodafone is a good Sun customer…. I am pretty sure Sun could make good money from Vodafone as an infrastructure supplier without the halo.

    What’s strange imho is the ongoing Java pathology, not Stephen’s omission of same.

    What would I do? Go on a new jihad – the storage is the computer. its all about the data and the i/o baby, cloud and on premise.

  6. @James I wasn’t talking about $$ wrt to Java – clearly other companies have done much better from Java than Sun in that respect.

    I was talking about brand value – Java is probably one of the few brands that people would recognize, it means something and has a huge ecosystem built around it – so it has value. Clearly little value to Sun but a some value externally. It would be sad to see “Java” become irrelevant due to Sun’s neglect / inability to extract value from it.

    – Rich

  7. @Rich Sharples: no conspiracy theories, just not much to comment on there. the client side, i covered, and as for the server side, i’d need a better picture of the revenue. what we have says that Java – all of it – is about a third of Sun’s software revenue. but which pieces of the Java portfolio – specifically – are making money? ME? GlassFish? what?

    given those answers, i’d better be able to comment on the overall future of Java. but clearly it’s an asset; i just don’t know how valuable.

    one addendum: it’s been said to me that NetBeans exists to service Java, and undoubtedly that’s true to a certain extent. but that – to me – is a problem solvable by other means.

    more on that later, maybe.

  8. @Matt Raible: agreed. there has been a lot of concern in the past, for example, with what would become of OO.o were Sun to force it to stand on its own two legs.

    my opinion has been that it would be fine, and even if not, it must be left to its fate. Sun cannot afford to fight the good fight any longer.

  9. Rich Sharples: “Clearly little value to Sun but a some value externally. It would be sad to see “Java” become irrelevant due to Sun’s neglect / inability to extract value from it.”

    Java relevancy is clearly not Sun’s shareholders problem… making money is the charm.

  10. have a look at

    Will explain why there is an intersection between amazon type cloud vendors and Big iron vendors like Sun

    Netbeans -is like Apache – Really a cost savings function no complementary sale. Why not get onto the eclipse bandwagon and get some real costs down.

    Openoffice – Does Sun think it can compete in thise consumer focussed space and compete with Google and MSFT ? Dellusional

    J2ME – With Android and the full stack approach – why should any cell manufacturer go with J2ME. The value of having a Java programming env is already accomplished

    At 3 dollars a share – maybe it is steal. But I said the same when it was 12 , 10 , 8 and 5 and now 3. Having put in $35K of my wealth into the company and dont have the stomach to put in more.

  11. Steve,
    Good write up, very thought provoking. No doubt there are big decisions that need to be made by management to ensure that the company is focused on technologies and products that can serve customers and drive revenues.

    I believe that management will do the right thing to get the company back on a growth path by focusing on good execution to back up the strategy. That may require doing things differently than the company has in the past, but that’s what’s needed.

    Your posting (and Tim’s) have already made the rounds quite widely at Sun.


  12. James: You’re starting to sound like normal analysts. Historically (1997-2002), Java was the vector through which Sun made all its money. It was the anchor for the relationships from which Sun’s sales force were able to sell the rest of Sun’s product portfolio to Sun’s core customers in the bubble years. Java made a respectable direct income for Sun (and actually that was part of the problem, more in a moment), but much more importantly created the context for Sun’s enormous success in the late 90s by giving the sales force a reason to relate. This was the whole reason IBM stopped promoting the Java brand in 1999-2000 – the relationship sell by Sun was seriously hitting their server sales in Europe. So saying Java is “a technology that never drove significant sales at Sun” is based on a reductionism I never expected from Redmonk. The problem is not engineering-led failure.

    Sun’s problem at the start of the decade was twofold: the siren call of direct revenue promoted the reification of the Java brand and the loss of understanding of the relationship sell’s value, and the end of the bubble meant it stopped being a vector to relationship sales anyway. Both led to the place we are today because there was no model to fall back on.

    Today the big question is: has Jonathan created a replacement model through his leadership? I believe he has, with open source as the key to the relationship. For me the question is whether Sun can bridge these years at the end of the bubble collapse he inherited, and that’s what the restructure is about: bridge without breaking the new model.

  13. The way out, I think, for Sun, is to virtualize the cloud. I think it can. What I mean by that is: you run Sun hardware at some company datacenter. Then you run similar Sun hardware in the third-party datacenter. The virtualized cloud allows you the luxury of moving the running application and storage between data centers, either on-premise or at “clouded, we don’t know in what city” datacenter while the application is running.

    (Take a look at what EVE Online did with Stackless Python for similar stuff)

    What to do with mySQL? make it autoshard and auto-unshard. The thing you need in the cloud that absolutely nobody has is a SQL 92 compliant database that scales to cloud levels. Sharding is not the perfect solution, but it’s better than reinventing SQL.

  14. SOG, minor quibble with some points, but otherwise +1.

    I think you take the RIA distinction too literally. IMO, Gmail/Docs is RIA. Which makes their cloud computing strategy even more critical. It’s a make or break battle for them (along with MSFT). New apps, new deployment model. RIA/Cloud are two sides of the same coin. However, you’re probably right about JavaFX….Finally, don’t think cloud computing margins need be razor thin.

    @Simon Phipps. Not sure a modern open source ‘relationship’ can ever recreate the Java relationship of ’99-00…

    Chris MarinoNovember 26, 2008 @ 2:06 pmReply
  15. On the RIA point, Tim and I disagree on this one (and likewise I think with SOG here …). The problem is two-fold. From a provider standpoint (pick a vendor with a storefront or website), while the browser makes it easy for me to set up shop, it also makes it really easy for a browser (or any elements, like a toolbar within it) to take my customers elsewhere, perhaps my competition.

    From a customer standpoint, yes … on my desktop, the browser provides a reasonably rich experience … however I have never (rarely?) had such a wonderful experience on any other device other than my desktop. Most browsers on mobile devices, well … ugh.

    Now maybe I’m being too literal here in terms of thinking of the browser (as opposed to the technologies within it, embedded within my apps on my device, ala WebKit), but that again leads me to point #1. Where I end up is that no, I don’t think the war is over yet, just that a number of battles have been fought and won in the space … but it’s a journey and I don’t think we’ve yet arrived.

    Robert BrewinNovember 26, 2008 @ 5:27 pmReply
  16. @Chris Marino: Maybe, maybe not, but that wasn’t my point; just batting down the tired canard that “Sun never made money from Java”. The problem is the opposite, that it made so much from Java that Java gained a special status.

  17. Good write up Stephen. I agree with most of what you wrote.

    I think Sun needs to reassess its open source policy of ‘no commercial software’. They need to be able to generate software license income from their install base. Your idea for network services is on the mark but it should also include selling value add commercial software.

  18. now now Simon no need to start with an ad hominem attack. :-) The money was in database servers and network servers. i don’t think you give Solaris nearly enough credit as a power brand for the late 90s. its also a bit unfair to accuse me of reductionism at this point given that my position has always been fairly consistent on this. I work closely enough with IBM to know how much money it has made *directly* from “Java” – and that’s before you kick in all the associated services.

  19. @Ian Skerrett: Do you have a link for that policy please?

  20. @James Governor: So you disagree with me?

  21. Simon,

    Here are statements made by Jonathan Schwartz. Not sure if they are policy or accurate but Tim O’Reilly is usually a pretty reliable source.

    Are you implying that Sun could/will make software available under only a commercial license?

  22. Stephen,

    Thanks for this. Here are some of my own thoughts:

    — Terrence

  23. @Ian Skerrett: MySQL already does (as you know) and as I indicated on my blog[1] monetisation between deploy and scale is open to all approaches. Between download and deploy, everything will be open source, yes.


  24. Stephen, What is your view of Sun’s Identity Manager product suite? Where does it stand in the Billings Disclosure? Thanks.

  25. Stephen: I think you’re right on that Sun’s problem is that it is technology-driven. Technology-driven isn’t all bad, but now isn’t the time for it. That probably means the Sun we know probably can’t survive as we know it. My thoughts are at

  26. As someone with 20+ years of software engineering experience, allow me to offer another way of looking at Java’s relevancy to Sun.

    Java is a language and platform. Without languages, no software can be built. Whether Sun makes money from Java is not the point here. Java is a necessary technology to differentiate Sun’s other technologies like Solaris and Glassfish because Java is a first-class citizen on them.

    Here are other examples of language/platform technologies that do not make money:
    1. Objective-C from Apple – required to write OSX and iPhone applications.
    2. .NET from Microsoft – suite of languages for next-generation apps for Windows.
    3. DirectX from Microsoft – used by virtually all PC-based and XBox games.
    4. ActionScript/Flex from Adobe – Flex SDK is free but the tool/IDE is not.

    I think the examples above are enough for all to see a pattern here. This is why Java is necessary for Sun, whether or not it makes money.

  27. Many good comments here, but I think your take on NetBeans misses the mark. NetBeans has evolved into an excellent tool that acts as a “gateway drug” to Sun’s other offerings. It certainly worked that way in my case, leading me to re-engage with Java and eventually investigate Solaris and Sun’s hardware offerings which were completely off my radar previously. Good developer tools are important to Sun’s business and I don’t think the open source community on its own can provide them. I don’t see the same fit with OpenOffice, as much as I like and use the product.

    Miles ClarkNovember 29, 2008 @ 12:00 pmReply
  28. What about becoming the preeminent Java stack?

    I’ve seen companies like Azul that do so.. why doesn’t Sun have “Java Mainframes”? Highly tuned machines that will handle tons of transactions and require no upfront tuning.

    They could do the same for PHP, Python and Ruby running on the same JVM.

    Imagine a system that has all the software it needs so that you hand it a WAR and it runs any application you give it. But it can run up to thousands of simultaneous threads on the same VM.

    Worried about security? Get your own “cloud computer” in a box.

Some HTML is OK

or, reply to this post via trackback.

Continuing the Discussion

  1. […] wrote an awesome long form post on the subject, so I don’t need do. He was responding to Tim. Here is what I would […]

  2. […] I’m piling on too, with a somewhat different take than Tim Bray, Stephen O’Grady, James Governor, and Rich Sharples who’ve all contributed their excellent insights into this […]

  3. […] What Would SOG Do?: Views on Sun Stephen O’Grady, RedMonk […]

  4. […] topic of Sun, we do the favorite parlor game of playing “what will happen to Sun.” See Stephen O’Grady’s excellent write-up on that topic, as mentioned, as […]

  5. […] that’s not Adobe, including Sun and Microsoft. But, it’s not too late. I’ll leave the portfolio analysis to Stephen’s comments on Sun, which I pretty much agree […]

  6. […] the deprecession. But independent of that, certainly focus was a problem. We’ve recommended in the past that Sun decommit from underperforming assets that have either zero or a negative return, and it […]

  7. […] has traditionally made its money, could be a significant distraction? A: Absolutely. I’ve argued in the past that one of Sun’s issues was a lack of focus, and trying to do too much, in too many areas, […]