Larry Augustin's Weblog

IBM's Response to Oracle + Sun: Buy SAP?

In my previous post I speculated that Oracle will sell off most of Sun's hardware businesses allowing them to keep Sun's software businesses effectively for free.  I still think that's the right strategy for Oracle.  It would allow them to cement relationships with partners like HP, Dell, Fujitsu, EMC, NEC, etc. while further isolating IBM.  If Oracle remains in the hardware business they are, like IBM, competing with everyone on that list.  This only serves to strengthen the Microsoft relationship with companies like HP and Dell and isolates Oracle.

But let's assume for the moment that Oracle does stay in the hardware business.  How does IBM respond?  IBM has traditionally positioned itself as an infrastructure and services company, drawing the dividing line in its businesses at applications.  IBM provides the infrastructure (hardware and software) and services; its partners provide the applications.  This has worked well in the past.  It has encouraged application developers to partner with IBM, building on IBM infrastructure and turning IBM into the leading application platform vendor.

Oracle, in contrast, started as a software infrastructure vendor and many years ago decided to compete with it's partners by moving into the applications business. As Oracle has gobbled up applications vendors, IBM has seen it's software partners shrink. With the acquisition of Sun, Oracle now competes squarely with IBM in infrastructure in both the hardware and software spaces. Add to this Oracle's already strong presence in applications and it presents a problem for IBM.

The downside to being in the infrastructure business is that you depend on the application to pull demand for the infrrastructure. That is, business people decide first what applications they need and second what platform they need for those applications, not the other way around. So if your applications portfolio (either directly or through partners) is strong, infrastructure does very well. If not, it's a very difficult business. Oracle today owns a healthy applications business that can pull its infrastructure. IBM does not.

Further, IBM's dividing line of applications was drawn at a time when there were more applications vendors who needed its infrastructure.  Today the landscape has thinned.  Companies that used to be infrastructure like Oracle and Microsoft have entered the applications business.  Microsoft, arguably IBM's largest competitor as a platform vendor, is in the applications business.  Finally, in another trend worthy of a separate discussion applications themselves are beginning to look more like infrastructure as they become platforms for third party extensions.   All of this leaves IBM looking lonelier and lonelier without applications as part of its portfolio.

Enter SAP. Given the competitive landscape an IBM acquisition of SAP no longer looks as threatening to the applications market as it once did.  Now it just looks like IBM trying to stay competitive with the likes of Oracle and Microsoft.  Further, SAP finds itself isolated as the infrastructure partners it needs (e.g. Oracle and Microsoft) increasingly compete for its business with their own offerings.  By acquiring SAP IBM solves its gap at the applications tier in the stack and remains competitive with Oracle and Microsoft.  By combining with IBM SAP gains the rest of the stack it needs to stay competitive in the middleware space.  It's a win-win for both parties.

Oracle Buys Java and MySQL for Free

Glyn Moody and Matt Asay have already written good articles on what Oracle + Sun means for Open Source.  What I think is brilliant about Oracle's move is that they get incredibly important technologies like Java and MySQL for free.

My prediction: Oracle will sell the hardware parts of Sun's business (storage, servers and SPARC) to some combination of Hitachi, EMC, Fujitsu and HP and more than return what they paid for Sun.  In the end they will get some great software assets from Sun for free.

Patience has proven the right strategy for Oracle.  Rumor is that Oracle offered as much as $850M for MySQL.  Oracle refused to bid up the price against Sun and Sun ended up paying $1B for MySQL.  But in the end Oracle gets the last laugh. By being patient they have managed to get MySQL, plus all of Sun's other software assets, for free.  Let's run some rough numbers and see how that math works.

$7.4B	Total purchase price
-$1.8B	Net cash at Sun
------
$5.6B	Cost after using Sun's cash

So we have to raise $5.6B to cover the cost of the acquisition.  For the fiscal year ended June 30, 2008 Sun reported $13.9B in revenues.  The fact that Sun had an Enterprise value of only 40% of revenue is an artifact of people not believing in the business.  It doesn't necessarily reflect the true value of those assets.  This is where Oracle gets a bargain.  In fiscal 2008 storage products represented 17% of revenue, or $2.4B.  Natural acquirers for that business include Hitachi, EMC and HP.  Using HP as an example, the company is trading at 1.4x revenue.  (Yes I know we should more accurately be looking at earnings as a better proxy for valuation, but these are rough numbers and detailed earnings numbers for Sun's business segments are not publicly available.) Sun's storage bushiness should net somewhere between 1x and 2x revenues.  Assuming the low end of 1x that leaves us with $3.2B.

Next, lets look at Sun's systems business.  According to SEC filings sun did $6.2B in computer systems products revenue in fiscal 2008.  Sun does not disclose software revenue separate from systems revenue, but industry analysts estimate about 15% of total revenue comes from software. Let's take the conservative approach and assume more of |sun's revenue is due to software, lowering the potential value of their hardware business.  At 20% of total revenue that's $2.78B, leaving $3.42B for the server and Sparc parts of Sun's business.  Again, taking a conservative 1x multiple of revenue we have:

$7.4B	Total purchase price
-$1.8B	Net cash at Sun
------
$5.6B	Cost after using Sun's cash
-$2.4B	Storage
-$3.4B	Servers and SPARC
------
-$0.2B	Profit for Oracle

So Oracle gets Java, MySQL, Solaris and Sun's other software assets for free.  Seems like a pretty good deal to me.

So what's next for Oracle?  The on-again, off-again rumor has been that Oracle will acquire Red Hat. Recall that Red Hat acquired JBoss for $350M.  Could Oracle get a bargain on JBoss by being patient yet again and letting Red Hat make that acquisition as well?  RHT is trading for $17.36/share with a market cap of $3.3B as of today.  When Red Hat acquired JBoss they were trading at $29.85 per share.  By being patient Oracle could be getting Red Hat and JBoss at a 43% discount.  Not bad.  Patience can be a virtue.

Chrome: Google’s New Operating System With Built-in Web Browser

Perhaps the most interesting question raised by Google’s Chrome is when does the Web browser stop being a Web browser and start being the operating system?  For those of us that have been in computing for 25+ years now it's fascinating to watch technologies come all the way back around. Since it's beginnings the Web browser has moved from document viewer to application UI to application platform. Can operating system be far away?

Pardon me for a moment while I digress a bit and show my age.  If you're not an old-time Unix user you may want to skip to the next paragraph.  How many of you old-time Unix users out there remember the jokes about the "Emacs operating system‘?  For those of you not familiar with it, Emacs started life as an editor with a complete Lisp interpreter built-in as a way to extend its functionality.  Of course with a complete Lisp environment built-in, people eventually developed some fairly complex functionality inside the system, including email clients and a Web Browser. People have even experimented with Emacs running standalone on a Linux kernel.  The idea being that the only application environment one really needed was Emacs.

With Chrome Google is taking the Web browser in exactly that direction: making it the application platform.  None of this should come as a surprise.  The more people use the Web, the more money Google makes.  That's the beauty of Google's business model. Google doesn't make money when you or I install and run a piece of software locally on our machines.  So Google has to make the Web application experience as good or better than the local application experience.  To do that they need to take a big step forward from the existing browser technology and re-architect the browser assuming that's the use case.  That means treating each "Web page" as a separate application, running in its own memory space, isolated from other applications.  It also means taking the application development language (Javascript) and dynamically compiling it down to machine code for best performance.

We are also not always connected, so applications need an offline framework (Gears) that always them to run on a local copy fo the data and synchronize when connected back on line.

Sounds to me an awful lot like an operating system.

Well not quite. Chrome doesn't manage hardware. And although it runs each application (Tab) as a separate task and manages those tasks it doesn't do the task scheduling, swapping and other low-level memory management features an operating system does.  So there's still some work to be done.

Oh yeah. It lets you view Web pages too. :-)

Where does Google go with this next?  How hard would it be to add the hardware and process management features of an operating system?  Well there are already several Open Source projects out there that do that. Linux and FreeBSD come to mind.  A Linux kernel that boots and execs Chrome can't be too far behind. Anyone want to bet on how quickly someone makes a minimal "boot to Chrome" Linux installation?

Search for ACM Distinguished Engineers

The Association for Computing Machinery (ACM) is searching for nominations for Distinguished Engineers.  Distinguished Engineer is a new ACM member grade recognizing those ACM members with at least 15 years of professional experience that have made significant accomplishments or achieved a significant impact on the computing field.  I've been an ACM member since 1985.  The ACM has traditionally had a greater focus on academic Computer Science.  It would be great to see more participation from and recognition of Open Source practitioners.  The creators of many of the core Open Source projects out there are certainly deserving of Distinguished Engineer status.  Please consider nominating your favorite Open Source guru.

AJAX VLAB Panel Discussion Recap

Last night's AJAX panel discussion hosted by the MIT/Stanford VLAB prompted some thoughts on where browser technologies are going.  One of the points I made in the discussion was that AJAX was just a technology for building the client side of a client-server application.  It's an important technology for doing so because it has achieved a ubiquity that other client-side application development technologies have failed to achieve.  What we've done is create an application that is ubiquitous (the Web browser), and that application has a way of downloading and running code within its own contained environment.

This is a very important point that I don't think most people get.  When you visit an AJAX enabled Web site, you are downloading and running an application on your PC.  That application just happens to sit within a browser, and does not require going through your operating system to install.

What this means is that Web browser programming model will continue to evolve to make the process of developing downloadable client-side applications that run within the browser easier.  The Web browser is a cross-platform UI/client platform.  How long before "Web" applications are stored and run locally from your Web browsers "cache" without the need for any Internet connectivity?  It's not that far off, and sounds a lot like a downloadable client, doesn't it?

MIT/Stanford Venture Lab: The Next Wave of Web 2.0

On Tuesday February 21 I'll be part of a panel session presented by the MIT/Stanford Venture Lab entitled 'The Next Wave of Web 2.0: Is Ajax Where the Action Is?'  The session will be  6-8:30 p.m. at Bishop Auditorium, Stanford University Graduate School of Business, 518 Memorial Way, South Building, Stanford.  Om Malik from Business 2.0 will lead the discussion.  Other panelists include Scott Dietzen from Zimbra,  Dan'l Lewin from Microsoft, and Pradeep Tagare from Intel Capital.  Here's the abstract:

VCs and entrepreneurs are talking about a "funding frenzy" in the AJAX powered applications space. What is AJAX? It is the special sauce behind a new generation of rich internet applications such as Google Maps, Yahoo's Flickr, and a plethora of emerging Web 2.0 applications.

While some technology pundits are dismissing AJAX as nothing but a cool technology that enables web applications to act like desktop applications, others are claiming that there is more to AJAX then enhanced user interfaces. Many of the emerging AJAX companies have adopted disruptive business models that have successfully displaced incumbents' offerings.

Come and hear several break-out companies, including Zimbra, which are on the cutting-edge of Web 2.0. The panel will explore the source of value in AJAX, the viability of new business models, and the impact on software and user experience.

Why Does Beta Last So Long?

There's an interesting article by David Kesmodel in the WSJ on how companies are now having longer and longer beta test cycles for software.  The article was also discussed here on Slashdot.

Kesmodel is critical of companies for long beta cycles and for turning the concept of extended testing among real users (beta-testing) into a marketing exercise.  I'm sympathetic to that criticism, but I think that since the terms "alpha" and "beta" came into use at IBM in the 60's to refer to test cycles the complexity of software has increased dramatically, and our ability to test that software has not.

Three major drivers of this phenonoman are 1) the number of users, 2) complexity of interaction with other sofwtare, and 3) the desire for early and broad user feedback.  First, it is virtually impossible to test major Web applications at scale.  If you're Google, how can you possibly replicate the deployment platform (tens of thousands of systems) and test the shear volume of traffic?  The only true test is to open it up to the real world.

Second, the number, variety, and configurations of systems that your software must interact with is truely mind-boggling.  How amny Linux distributions are there?  How many Windows registry variations are there?  Even trying to test all variants of Microsoft IE alone can be overwhelming.  You have no choice but to open up testing to the world.

Finally, "release early, release often" has become a standard mantra (for good reason) among developers today.  Successful Open Source developers practice it religiously.  What's the equivalent in the proprietary software world?  The extended open beta test.  How long was it before Linux 1.0 was released?  Three years?  Proprietary developers are just doing the same thing as their Open Source counterparts - getting early feedback before all of the features have been implemented.