Oracle had it’s big, 5 hour event today explaining what they’ll be doing with Sun and the new strategies and company that will result. As immense as the topic is, below is an equally sprawling selection of my commentary. While folks like myself relish this kind of thing, I expect most others will have what they call a “meh” reaction, not too sure about what’s new and different. Well, Oracle trying to be your single source for all enterprise IT is the main thing, Oracle trying to take over IBM’s market-spot as a second. More details below, starting with a re-counting of my tongue-in-cheek summaries from Twitter.
Summary a la Twitter
- You’ll get one big SPARC box, that does everything, and it’ll be f’ing awesome.
- How are we going to do better than Sun could? Spend more money & ABC. Coffee is for closers.
- “We don’t integrate, we eliminate.”
- We still have a lot of middleware, MySQL will separate org, OracleVM + SunRay + ERP = new sales strategy.
- Enterprise Manager/OpsCenter BFFs. Sun/ORCL IdM is more about about who gets the kids.
- Our support is going to blow you away; our core business is fattening margins. Lunch break!
- Ah-hah! Larry says “cluster” and “cloud” synonym, and Solaris is the bestest for running “clusters.”
- “It’s great that Hasso and his five guys got it. It’s whacko!” –Larry Ellison on in-memory databases.
Summary: Portfolio Rollerball
The future of Oracle and Sun are fully integrated stacks, from metal-to-glass. While many were expecting – or even gunning for – a portfolio blood-bath, Oracle seems to be keeping most everything, with only parts of Sun’s identity management suite falling under the “we’ll be integrating that technology” block. Other portfolio “losers” simply got demoted to niche status, like NetBeans and Glassfish whose status will pale in comparison to the “enterprise” ribbon given to JDeveloper and WebLogic.
Tech-industry folks like other vendors and partners have the most to be concerned with in the here and now – figuring out how they’re going to be with ‘em or agin’ ‘em. IT buyers and others have to ask themselves if they want the single-stack, “one throat to choke” model of if they’d prefer hedging with heterogeneity.
The Single Stack
“Our vision for the year of 2010 is the same as IBM’s for 1960… That strategy, by the way, made IBM the most important [technology?] company in the history of the earth.” –Larry Ellison
As most people presumed, the Oracle + Sun strategy is to offer the best durn’ed stack from server to OS to middleware to applications and even to end-user device. There’s the Sun server, storage, and networking gear starting from the metal; Solaris and Linux, all virtualized of course; databases with Oracle DB, MySQL, and more; Java middleware; identity and systems management; enterprise applications; and even SunRays to get all the way to the end-user’s eyeballs and finger-tips.
“Vertical integration,” as Larry Ellison put it. To be clear, in the wrap up Ellison said they’d still be happy to sell components, Oracle DB running on HP, he said. It’s the vision of 1960’s IBM with one stack, but the sales approach of 2000’s IBM: use whatever technology you like, so long as you buy it from us.
Why Oracle thinks it’ll work this time
At Sun events we’d have AC/DC cover bands. At Oracle events they get AC/DC. –Craig “@ThinGuy” Bender
If this was to be the strategy, the biggest question for me was how Oracle would succeed with the “we do it all” strategy where Sun failed. Sun’s dream was largely along these lines – most recently manifested in what I called “The Pink Dot Theory” for the world map full of pink dots denoting Sun IP that Sun CEO Jonathan Schwartz would often show. The answer is a combination of spending more money and optimizing, including changing from a build-to-stock to a build-to-order scheme for their high-end hardware.
A major part of that is being more high-touch, as Ellison put it:
We just have to do a better job [than Sun did] taking this [excellent Sun engineering] and delivering it to customers directly.
As such, much of the talk in the 5 hour event touched on services, supply-chain fixes, focusing on high-end vs. commodity offerings, and in general working more directly with customers.
The single-sourced stack
This is the grand old question of tech companies. Technology companies love to claim that customers want “one throat to choke,” one vendor to help them simplify all the complexity, and so on. To vendors, it means all parts of the stack to get paid for.
The good side, as highlighted by customers on the stage time and time again was a reduction in complexity: having to worry about integrating less things, having to manage less things, etc. It’s kind of the same appeal of the cloud: less stuff.
The bad side of having one throat are the risks: IT buyers are afraid of single stacks because it locks them into price-hikes, risks of technology slow-downs, and otherwise presents the risk of harming them in the long-term. Open source was used as a response to these risks when manifested in the past 20 years.
It’s interesting to think of other parts of the tech world where “single stacks” are sold: networking, performance enhancement pizza boxes, mobile devices, and so on. Most any category – which is most of them – that aren’t “open” have single stack approaches: servers and enterprise software are one of the few places that can be – and are – decomposed into parts from various vendors.
There wasn’t much mention of “cloud computing,” with words like “clusters” and “grids” being used as synonyms. Larry Elision reserved most of the cloud zingers for himself, and they amounted to one of his kinder statements “everything’s called cloud now, there’s nothing but cloud computing.” Which is another way of saying, if everything is a cloud, “cloud” is meaningless.
And, of course, they took what I call the LL Cool J stance on cloud computing: don’t call it an innovation, we been doin’ it for years! With virtualization and cloud computing, this is what every elder company has said, so it’s expected.
Ellison said Solaris would be the best at running clusters, “or I guess the [popular] word is ‘cloud,'” he tacked on. So, there you go: cloud for Oracle is about all that existing cluster and grid stuff we’ve had for year. Presumably, with new innovation laced on-top. And really, with “cloud computing” potentially having the same specter as “open source” (cheap), who can blame that attitude?
The more interesting question for Oracle is answering the call of SaaS, where “SaaS” means “paying less and having an easier time managing the applications because I don’t actually manage them.” I’m not sure there’s a clear answer from Oracle on that one at the moment except, “you probably don’t want that.”
There was much more than this broad strategy. Here are some collected items:
- x86 – We’re not interested in commodity, x86 market, Oracle’s Charlie Phillips said. There’s other people who do that, like Dell or “whatever,” as he put it. Instead, Oracle wants to focus on high priced hardware, just as Sun did. There are x86 offerings, John Fowler added later, but they’ll focus more on developing enterprise class functionality for Linux. “OpenSolaris” wasn’t uttered once, it seemed.
- Java – When it comes to the core parts of Java, things will pretty much stay the same, continuing with: more modularity, glomming on dynamic languages, multi-core support, and making it faster.
- App servers – WebLogic will be the enterprise solution, Glassfish will be it’s little brother, for “departmental” needs (not top on the salesforce list). (This pattern was applied to much of the Sun portfolio.)
- JDeveloper and NetBeans have little overlap, according to Oracle’s Thomas Kurian. JDeveloper will continue to be for general enterprise development, while NetBeans will be niched for mobile, dynamic languages, and visual-oriented development.
- Identity Management – you’re going to need a diagram to figure out how Oracle and Sun’s identity, directory, and security portfolio shakes out. The upshot is that they’re niching Sun’s directory down to telcos and other places where “people want directory on-top of file system.” Most of the identity stuff will be folded into Oracle’s existing offerings, but the Sun Role Manager seems to have “won” in it’s spot. Hopefully OpenSSO will land well: they’ve been doing well keeping up with emerging identity technologies from the consumer world.
- SOA – Oracle’s offerings “win,” but there’ll still be support for Java CAPS and OpenESB, for integration at least.
- IT Management – Sun’s OpsManager is one of the more notable “survivors.” Oracle’s own IT Management offering, Enterprise Manager and OpsManager are “very complimentary to each other,” so it seems they’ll be happily combined. The larger question here in light of the single-stack approach will be – as it always was for those two – how well they can be used as a general IT Management suite. That is, while they may be the “best” way to manage Oracle + Sun offerings, how will they do at managing other’s technologies, like Exchange, SAP, EC2, etc.?
- OpenOffice – Oracle’s Edward Screven used the phrase “Cloud Office” in reference to OpenOffice, which sounded nifty. Need more drill down on that one.
- Virtualization will be OracleVM (which is free), which will work with Solaris now. He had good things to say about containers/LDOMs in Solaris.
- VDI – With SunRays and (to a lesser extend) VirtualBox, Oracle can now have a desktop virtualization strategy which Screven was actually pretty bullish about. I haven’t heard so much SunRay love in a long time, even from Sun folks. SunRays could be a nice boat if there’s a rising tide of VDI lust.
- Partners – When it came to partners, the message was clear: add value or go away. Oracle’s strategy was to be more direct with customers than Sun had been, which reins in the slack that partners could live off. Partners in the Sun ecosystem, then, need to make sure they’re actually adding something unique rather than taking advantage of inefficiencies in the ecosystem.
- Consumer-tech – Sun had been pursing consumer technologies for awhile, and the exact direction Oracle will take was subtly unclear. While Elision said something along the lines of “we’ll stay in the data center, won’t get into consumer electronics,” there were still mentions of being “arms dealers” for consumer technology back-ends and tools. For example, the emphasis on NetBeans as being used for visual work, or Java being available on millions of handsets. You could see how Oracle would look at Sun’s Java Store as a framework – cloud “middleware,” if you will – for all these companies who now want app stores. And there’s a lot of people like that: during the break I spoke with someone who’s getting into that, and I talk with companies large and small about app stores almost weekly now.
What it means for competitors & co.
This is a big moment for the IT industry. All of us have been waiting to see what’d finally happen, and this shifting re-creation of Oracle as a single-stack provider is new, unnerving many of the long-standing partnerships and strategies for other vendors.
The partner and segment musical chair chance folks like VMWare, Cisco, HP, and Microsoft have been making seem warranted now. IBM’s attempts to up-level their role in the world to “solutions” instead of just technology is interesting as well. Dell continues to be a weird position in all this: Oracle’s fine for them to pick up the low-end stack, but Dell’s ambitions to get into the high-end now have another tough-as-nails competitor. Mobile folks will probably have it easier without having to worry too much about Java competition, or find it easier to work with. While this may seem like a relief for Adobe Flash and Microsoft Silverlight – less competition from JavaFX – those two have a bigger fight in progress from the “open web” crowd.
One of the more interesting winners seem to be the Big Data, in-memory database crowd. Larry Ellison dismissed much of that as something they already do well or a technological pipe-dream, depending on the technology, buying innovators in this area time to innovate and build markets without too much Oracle defensive reacting from Oracle who doesn’t seem to take them seriously. If there really is something new there – even if it’s just making it cheaper – then there’s a tiny window to get Oracle with their pants down.
Partners seem to be in the most desperate position: they need to figure out how they can prove their worth to Oracle.
What it means for IT & co.
For IT folks, the big question revolves around your desire to become a dedicated Oracle customer. It may only be high-end enterprises that need worry themselves with the larger question here – as was made clear, Oracle is still happy to sell piece-mail instead of cater your whole digital meal.
As detailed above on “cloud, and as I Tweeted, there wasn’t much talk of “as-a-Service,” and the focus was still “as-a-SKU.” When talking with Oracle about cloud or SaaS, the first response you’re going to get is probably “you don’t actually need that, and let me tell you why.” While they may actually have offerings, Oracle is at the stage IBM, Microsoft, and other were 1-2 years ago: they won’t play along with cloud reframing and buzzword-bingo, believing that there’s nothing new there. Notably – for better or worse – Oracle is one of the last elder companies that does this.
Unless you have need and budget for high-performance hardware and middle-ware, it’s unclear if Oracle will care (revenue-wise) about the LAMP-stack, open source, “lesscode” crowd. That said, Ellison made a comforting comment about Java: it doesn’t have to create direct revenue, instead they’re concerned with it helping to enable revenue for the whole pie, not it’s little slice. Oracle clearly believes that its closed source offerings (Oracle DB, WebLogic, it’s apps) are “better” than open source alternatives (MySQL, Glassfish, etc.), but as long as the open source ones stay in there place, it seems they’ll be copacetic.
Updated as I come across more:
- As Simon Phipps put it, “[a]ll the dirt is in what hasn’t been said.” Stephen and James are often better at delivering that dirt than I, so keep your peepers on their blogs for that.
- NY Times DealBook piece.
- Judith Hurwitz asks five questions.
- Tony Baer’s commentary.
- Joab Jackson’s piece narrows in on virtualization.
- Vinnie Mirchandani’s commentary on price hikes ahead.
- Gavin Clarke rounds up the open source keepers and slashers.
- Paul Krill gets a definitive quote on Sun Cloud: “We’re not going to be offering the Sun Cloud service,” said Edward Screven.
- The New York Times’ Ashlee Vance has two pieces: a blog entry that wins the prize for using “quixotic” and a straight up article on the topic.
- A nice roll-up of several analyst comments and her own from Loraine Lawson.
Disclosure: most every RedMonk client has some skin in this game, so check the RedMonk client list for my potential biases. Oracle is not a client, though Sun was.