As you must have heard, Oracle is buying Sun. A deal this big has a lot going on. Really, the only thing you can do is wait and see what Oracle will do with the massive portfolio they’re buying. If you don’t like waiting, here’s some scenario-speculation, lacing in some fine speculatin’ from others as well – nothing beats the grasshoppers (this one included) out of the grass like a mega-acquistion. Some of these ideas and scenarios seem good to me, others seem tragic, but I’ve tried to lay them out with an eye towards reality rather than what I’d like to see happen.
The Viking Scenario
The first thing you can expect from a Oracle acquisition is due-diligence of the assets and a comparative analysis where Oracle has competing assets. Oracle will weigh up what’s worth keeping and jettison the rest. The latter will be marked by end-of-lifing via support and maintenance, or releasing code to the community – where it will fade and die. –Gavin Clarke
First, there’s the easiest scenario. Under this scenario, you can pretty much ignore the more nuanced and tweedy discussion that follows in the rest of this note. The simplest scenario is Oracle strip-mining Sun, selling off chunks that Oracle doesn’t need or understand (hardware), tanking the “unsellable” stuff, and keeping a few, key software assets (Java, Solaris, Identity Management, maybe MySQL, and a handful of others).
A large part of this would be a defensive move to prevent others (read: IBM) from acquiring and benefiting from Sun’s technologies. It’d also be one of the biggest Oracle chest-thumpings ever, like a lawn mower running over a frog.
The primary break-up here would be Oracle selling off Sun’s hardware business. Becoming a software and hardware company is difficult (ask HP and Cisco, or even IBM who makes it look graceful from the outside).
The point is that Oracle always has been and is a software company. Software people think differently than hardware people, and hardware people think differently than software people: one has to always be more important than the other, and it’s rare that one company can treat them as equals. As an example by anecdote, just look at the software that comes with your printers and scanners: no respecting software person would let that crap escape their fingers.
If Oracle is going to take Sun on as a whole, they’ll have to become one of those few companies that treats hardware and software as equals. The other problem with taking on Sun as a whole, as alluded to in the next first section, is filling out the Oracle consulting and services arm. Taking on Sun as a while would be risky if Oracle wasn’t trying to become IBM. To do that, they’ll need armies of consultants, which seems to point towards acquiring a services firm, as IBM did long ago.
Looking at Oracle past acquisitions, their initial, stated tendency is to speak towards fusing their existing software and newly acquired software together. That scenario isn’t too pragmatic if you can manage to extract good maintenance fees from the existing software (as Oracle seems to be doing with it’s enterprise software portfolio): rather than spend the money to integrate “legacy” software into the future, you just slowly keep ratcheting up the maintenance cost until your customers can’t take it any more, then you sell them a brand new thing that comes with the initial revenue bump of a rip-n-replace install. Hey, those yachts don’t pay for themselves, brother.
That’s the easiest, “worse case” scenario for the people at Sun: either your product produced revenue in the here and now (you stay), or you’re out. That said, as Gavin gets to in another part of his piece, Oracle is known for applying that same rule to itself: if an item from the Sun portfolio is better, they’re prone to throw the weaker Oracle item over-board.
But that’s the Viking scenario. What’s the more happy-path one(s)?
One Pair of Hands to Choke Many Throats
There’s one grand-theory put forth on the press release and the short conference call this morning: Oracle can build a “drive to application,” metal-to-screen stack of enterprise software. Previously, Oracle really only had middleware and enterprise applications to sell. Sun brings hardware, both on the server side and client side (with SunRays). Along with the ERP applications Oracle has, this means a business could buy “everything” from one vendor.
In reality, enterprise buyers like to buy across multiple vendors and also end up accreting all sorts of technologies and vendors through M&A and different CIOs. Less grandiose than buying “The Oracle Stack” is Oracle’s potential to sell to more parts of the enterprise IT stack, software and hardware.
I have a blind spot when it comes to the services & consulting arm that Oracle + Sun would have – it doesn’t seem like there really would be a large one. Enterprises buy 3 things in IT: hardware, software, and consultants. IBM has offerings in all three of these; that’s one of IBM’s strengths. If Oracle + Sun acquires a large consulting arm, you can start to see IBM business model duplication going on, which isn’t a bad business to be in.
This move would be the final stage of Oracle trying to evolve from it’s database identity. While there are still plenty of sales and places for the Oracle DB, MySQL, Postgres, and now cloud storage/databases layered with things like Hadoop are showing that, eventually, data management will be commoditized. Or put, another way, the length of time that a company can extract revenue from data management software has shortened. Oracle has long expanded its portfolio beyond the database with acquisitions like PeopleSoft and Siebel, and acquiring the enterprise IT portfolio that Sun has pushes Oracle into being a fuller enterprise vendor, lacking a services arm to collect that extra $8 for every $1 of IT-spend (or whatever kookie multiplier you want to use).
This perspective is from the vendor’s angle, of course, which means extracting more money from customers. To get a peak at the potential negatives from the buyer perspective, check out Howlett:
Who worse to entrust your entire stack than Oracle, that voracious consolidator of application providers and now, it seems, guzzler of open source and hardware?
He goes on to talk about body parts in vices and maintenance fees.
Surely Oracle will want to derive all the strategic value they can from the acquisition and subsequent control of the JCP. I can’t imagine that they would be willing to give up that control of the JCP in favor of a more open organization.
That’s a shame as I think all parties would be served in the long run by having a truly open Java run by a truly open organization. I’d love to see Oracle give up the reserved seats and allow the JCP to be truly democratic (or as democratic as any standards organization run by companies with competitive interests can be) but I just don’t see that happening.
In short, chances are low that the Java world will change for two reasons:
- Java has been an independent enough entity for long enough that screwing it up is difficult to do. Put another way, it’s BigCo interests have been mucking around with the Java world for so long that it can’t really get worse.
- Java is at the heart of much of Oracle’s software, and its at the heart of Oracle competitors like SAP and IBM. Java is so inter-woven into the enterprise software industry as a whole (on both the vendor and buyer/enterprise side), that it’s hard to screw around with it. When innovation in the Java world slowed down, JBoss and Spring came about to make up for that calcified process.
Not to mention that it’s open sourced – TCK, kiosk-use, and trademark shenanigans aside. What could be negatively effected, as Alex Miller gets to above, is a better future, sooner rather than later, for Java.
Screwing around with the status quo of Java has the potential to mess-up Oracle as much it would mess up it’s competitors. To put it another way, Java is already as screwed up as it can be. For as old as a technology as it is, it’s screwed up very gracefully, which should be taken as a compliment.
The worse case scenario is that Oracle won’t do anything with Java. They’ll let it stay the same, with some bullying in the JCP to very mildly serve Oracle’s interests. But, again, since Java is at the heart of so much of Oracle software and of their customer’s software, I’m doubtful that they’ll try to change it or calcify much more than already has. When something works in technology, you tend not to screw with it: you try to do the exact opposite, and keep it how it is as much as possible.
In the Sun part of the Java-world, there are some weird, from Oracle’s perspective, items to think on: JavaCards, JavaME (what does Oracle care about the mobile and embedded market?), and as explained more below, JavaFX.
“I’m an Oracle developer…er…hrm…”
Unless Oracle keeps Java at the same arms-length that Sun has, relative to the rest of the company’s portfolio, the Java community will have something of a identity crises.
Java programmers have always been, well, Java programmers, not Sun programmers. Netbeans bundling aside, Java has remained a rather separate entity in Sun-land. If Oracle messes with that sort of approach – for example, marging JavaOne into Oracle OpenWorld (making OpenWorld, perhaps, the largest developer conference) – Java programmers will be in a weird, new head-space they haven’t been in yet: alligence to a vendor, not a platform.
Redmonk analyst Stephen O’Grady noted important differences of opinion between McNealy and Ellison regarding open source that also “could present a problem.”
O’Grady told me this “will definitely be a painful acquisition… I don’t see any of these integrations happening quickly. There’s a lot of moving pieces” that have to be reconciled. *
The most repeated themes are around what Oracle would do with the open source part, essence even of Sun. As I told one reporter today, the open source community has been “mildly gloomy” today. Oracle has not established a fantastic reputation for some in the open source world, for example, coming out with Unbreakable Linux and just being a closed source company in general. That said, like all Elder Companies (those large, successful companies that’ve been around “forever”), Oracle has warmed up to at least using source if not promulgating business models around it.
In buying Sun, Oracle has to know what it’s getting into with the open source world: namely, a lot of it. As Stephen alludes to in the quote above, the cultural fit is not exactly fantastic, as good as it could have been with IBM. To use an excellent point I’ve heard from Sam Ramji a few times recently, those IBM lawyers understand open source, and that’s your first barrier to doing open source at a large company. Oracle’s use of Java and Eclipse shows that they’re a shrewd users of open source. Taking on the role of a being steward of open source, as Sun aspires to, is up in the air.
Large vendor involvement in open source is largely along the lines of funding and budget for the full-time staff who work on open source. Oracle has touted the Linux kernel development it does and they picked up a smattering of OSS Java involvement with BEA. Still, Oracle is no where near as “involved” in open source as Sun, IBM, or others (I realize I’m setting myself for an exciting [re-]overview briefing from Oracle here on all their OSS involvement ;>).
We like to sell software, not give it away
I’m not really that rosy on Oracle believing in the idea of open source business models. Clearly, they’re convinced that using open source is good for their business, but selling open source, not so much as Sun.
This isn’t really all gloom and doom though. Sure, Oracle could layoff the Sun staff that dedicates the majority of it’s time to open source projects (you can bet they’ll keep up the clip of the existing Sun layoffs, if not more). Or, in that specter’s perspective, people who do not directly contribute to revenue generating software. But this has been the problem with open source of late: nailing the BigCo business model.
As gets covered quarterly by the usual suspects (myself included), open source has been proven to work extremely well to boot-strap a business (and for fantastic exits for the likes of the MySQL team), but questions remain about using pure open source business models. As I’m fond of quipping now-a-days, the way you make money with open source is to sell closed source.
As a company, I expect Oracle to get rid of parts of it’s portfolio (open and closed) that don’t make money directly, are on the path to making money, or provide part of the stack that money-making software runs on (see Ellison’s comments about Linux being the #2 operating system, after Solaris, that Oracle software runs on).
From Oracle’s perspective, as is the case with IBM, if you become a systems seller, open source is just a necessary “evil” for that more complete, hybrid open/closed + hardware + services sale you try to make. Some software you charge for, some you don’t, but the whole package you damn well charge for.
Today our servers are running various versions of MySQL, tomorrow they’ll be running the same thing, and if need be ten years from now they can run the exact some software. Because of the GPL every WordPress user in the world is protected — we’re not beholden to any one company, only to what works best for us. Today that’s MySQL, tomorrow that’s MySQL, a year from now we’ll see. –Matt Mullenweg
On the open source angle, everyone’s specific question is what will become of MySQL. For me, the answer here is that anything that was going to happen with MySQL already has. Much of the original team has already left (hey, why would you stick around once you were rich?), the community has already gone through a large company “acquiring” it, and Oracle has already figured out how to cope in a world where excellent, high-quality, high-volume capability database are free for the taking (they move up-market and generalize their portfolio beyond the database, see above).
Still, the question remains open: will Oracle somehow spike MySQL? For one thing, by the nature of being an open source project – trade-mark aside – that’s impossible to do, technically. The mythical “The Community,” could fork the code and go along it’s marry way. Indeed, the primary way to gage the overly general “The Community’s” reaction will be see if, when, and with how much glee someone forks the MySQL code-base this week.
More pragmatically, Oracle would be ill-adviced to mess with MySQL. While it certainly damaged database sales, MySQL is used in scenarios that Oracle databases would be ill-suited in. All of those LAMP applications running around would be a lot more tedious if the only option was Oracle, MS-SQL, DB2, and other closed source, large databases. As with Sun, the hope here is that MySQL can be two things: a way of keeping in touch with the (from Oracle’s perspective) lower market segments and, thus, an entry into accounts that would never think to talk to Oracle.
Taking a page from what most open source startups are doing now, Oracle would be well served to start wrapping closed source services and code around MySQL, building out some nice metallurgical bands of payment, all the way from lead (free) to Red (Oracle DB).
Of course, there’d be a certain Larry Ellison panache if they just chucked MySQL as a way of saying, “that was a silly way to for Sun to spend a billion dollars.”
There are a few interesting other items to consider:
- Cloud Computing – Sun has been working on its own cloud for awhile. They have something there as demo’ed of late. I’m not really sure if Oracle has put much of foot in the cloud computing world, nascent and hype-driven as it is now. Oracle hasn’t been too hot on the SaaS space (yes, aside from Larry’s investments in NetSuite and Salesforce), they’ve seemed to stick to the notion of selling on-premise licenses, like their SAP buddies. The closet to cloud that I see Oracle getting (I’m sure, in an “unofficial,” “these are my words not my employer’s” capacity) are Monsieur Vambenepe’s cloud meets ITSM musings, which are quite excellent.
- Identity – both Oracle and Sun tend to bubble up towards the top of the identity management space. Indeed, Sun IdM is a money-maker. I’d suspect some consolidation here as having two different identity management products in one house doesn’t seem to happen that often – compared to keeping PeopleSoft, Siebel, and Oracle Apps separate to keep their customer bases happy. I’ll even throw a bonus, irresponsible, whacky possibility out: maybe the UnboundID team will be back working with their pals if their high volume directory technology pans out.
- IT Management – Oracle has always provided management tools for it’s own software and stacks. They’ve never been a strong, general IT management provider. Sun hasn’t either, but before the virtualization wave, Sun seemed to be going for it with OpsCenter. Since then, virtualization – with their own xVM hyper-visor – has been of more interest to Sun. xVM is an interesting asset to watch Oracle sort out: they’ll have to figure out if they’d rather go with more Linux-traditional virtualization silos (Xen, KVM), stick with xVM, or just deal with VMWare and Hyper-V. Before even that, Oracle has to figure out if they care about hyper-visors in the first place. At first glance, I’m not sure they would. There’s also some room for automation software, a la BMC/BladeLogic and Cisco: do a search-and-replace on Cisco/Oracle over here and you’ve got your day-dream.
- JavaFX – thus far, in the RIA world, Oracle has had much interest in Flex. I tend to view JavaFX as re-write of Java GUIs, but Sun has emphasized the RIA angle on it, esp. when it comes to a front-end for non-traditional “screens” like phones, TVs, and other “embedded” devices. Frankly, I’m not sure what Oracle would make of JavaFX; my feeling is that they’d favor Ajax and Flex, but there is the (small) chance that JavaFX would be seen as a way to in-house rich user interfaces. The problem there is that distributing the Flash Player doesn’t cost Oracle anything – it doesn’t eat into their software margins; Oracle only spends money on Adobe tools if they want to use Flex. They’ll be buying CS4 whether they deliver rich UIs in JavaFX, Flex, Silverlight, or anything. Thus, the only advantage to in-housing an RIA technology is the strategic one of owning more of your own stack – it’s the same reason vendors like Oracle and IBM have their own Java VMs. Oracle might be interested in the video aspects of JavaFX – more data to stick in databases – but that seems like a stretch. Here, JavaFX’s association with Sun rather than Java is a weakness. If JavaFX was part of Java, then it’d be just as safe as any other Java item.
- NetBeans – like JavaFX, NetBeans is in a tight spot under the Oracle roof. NetBeans has turned into a fine, featureful product. Having been one of the “why is Sun wasting time on this in the face of Eclipse” chorus for many years, RedMonk typically has nothing against NetBeans now-a-days. Oracle is very much an Eclipse shop, and Eclipse-people tend to have little tolerance for NetBeans. At best, I’d suspect that Oracle would cherry pick technologies they liked from NetBeans – the dynamic language editors, perhaps the modeling tools – and phase out the rest. At worse, they’d just tank it straight-out in favor of whatever the Eclipse community provides.
- [Star|Open] Office – people who follow me closely know that I’m not a huge fan of (personally using) Open Office. That said, it is a great asset when it comes to having a stick to poke Microsoft with, something Oracle (and every Elder Company) relishes doing. I tend to fall into the camp of spinning Open Office into an independent foundation along the lines of Eclipse. The efforts to fully commoditize the “document stack,” in the way that Eclipse commoditized the IDE stack, have been constrained by cross-corporate conspiracy theories and “politics” (read: “I don’t like them, so we’re not going to do it”). Fragmentation in the “Office” space helps no one, having one vendor in that space doesn’t help anyone either: having one industry standard is what we need.
So…is this all good…or bad?
With all this, and other, considerations and scenarios on the table, it’s natural to wonder if this is good for the industry over-all.
Clearly, it’s good for Oracle. At worse, they can strip-mine Sun for a few key technologies and (no doubt) have some fun being the end-of-the-line for a Silicon Valley legend. At best, Oracle can acquire a services arm and become IBM.
The industry as a whole has three short-term risks to wade through:
- Oracle messing up Java, causing other ISVs and customers to look towards other platforms.
- Oracle messing up MySQL, causing countless web sites to look towards using another database.
- Oracle jacking up maintenance and (Java) license prices on existing Sun customers.
If 2-3 of those risks remain risks instead of turning into realities, things will be OK for the industry. I’m not sure if the innovation engine that was Sun engineering will be kept up under Oracle’s roof, but the cream of those existing innovations will at least have a home.
For even more thoughts from RedMonk, check out Stephen’s fine note on the topic.
The above is written without much mention of all the folks who’ll be caught in the middle of change, good or bad. That is, laid off. At RedMonk, I’ve gotten to know many fine people at Sun, and a handful at Oracle (just a handful because we do tons of work with Sun, not too much with Oracle). Tragically, people will be let go and looking around for new jobs. There’s no sort of glee or ice-cold analysis to be had at that, and, in fact, it sucks.
Disclosure: Sun is a client, so I suppose Oracle will be. IBM, The Eclipse Foundation, and Microsoft are clients as well.