« January 2008 | Main | March 2008 »

February 2008

February 29, 2008

February 2008 posts that appeared on Research 2.0

NOTE: The following blog posts appeared originally at www.research2zero.com and were later syndicated to Seeking Alpha and elsewhere

Server operating system indicator says both Windows, Linux gains continue
Posted on February 29, 2008

The year-end IDC Worldwide Quarterly Server Tracker factory revenue numbers were released February 27. From an operating system perspective, there’s really no new news in them (but see the press release or the relevant IDC report for vendor and other characteristics). But we lay them out there for comparison with the quarter by quarter results we blogged on in November, September and May 2007.
For 2007 vs. 2006, the net-net is Microsoft (MSFT) Windows-based systems continued to gain share primarily at the expense of “Other” (primarily legacy mainframe-operating-system based). Unix/Linux-based server revenue was basically flat year over year with Linux open source systems continuing to displace UNIX systems as expected.

Linux gained slightly more than a percent of share while UNIX lost just under a percent of share (not shown).
As we noted in November, it appears from eight quarters of IDC data that the Unix/Linux share of the market is stabilizing in the 40-45% range, although they hit 46% in quarter 4 2007 only. We will need to watch to see if this is an upward trend or a seasonal issue.

Although overall Windows-based server revenues are gaining at the expense of “Other,” such as IBM (IBM) mainframe operating softwre, the trends went the other way in quarter 4, so that is something else we will watch. I believe that blip was totally seasonal because end of calendar year buying is a 40-year pattern for legacy systems.

— Dennis Byron
Tags: Microsoft, open source, Windows, Linux, IBM

Microsoft keeps trying to put its open source background in the background
Posted on February 21, 2008

As we’ve been noting since mid-2007, Microsoft (MSFT) is making every effort to put its anti-open-source baggage behind it. This is covered in detail in our annual Microsoft report released in December. On February 21, Microsoft announced sweeping open-source interoperability “principles” related to its volume software products (Windows Vista, the .NET Framework, Windows Server 2008, SQL Server 2008, Office 2007, Exchange Server 2007, and Office SharePoint Server 2007) that basically put its agreements with the European Union Competitive Commission, announced in October 2007, into Microspeak.
The announcement has three implications:

First, previous tactical opposition to open source software (OSS) has been a distraction to Microsoft’s “Software Plus Service” strategy, which hopefully will become more about providing IT and business services than mundame closed or open technology terms and conditions. This means “Software Plus Service” is misnamed (but don’t get hung up on words, as the U.S. presidential candidates are saying to each other). The Software Plus Service strategy has been a work in progress since Ray Ozzie joined Microsoft and dropping all the anti-OSS tactics makes that clearer to investors.

Second, although at the February 21 press conference Microsoft specifically said this announcement had nothing to do with the proposed Yahoo (YHOO) acquisition, that acquisition has everything to do with bringing Microsoft services to consumers just as most of the current available Microsoft “Live” services support enterprises. Microsoft is uniquely positioned to support both enterprises and consumers. More importantly, it can support each individual in his or her enterprise and consumer roles as those roles change during the day.

Third, the announcement covers all interfaces used by Microsoft itself in tying its volume products to “other Microsoft products.” That means it will be easier for open source software (OSS) providers to connect to the BizTalk integration engine (if that’s considered separate from Windows Server 2008), the Greats Plains heritage application software (even the smallest OSS ERP provider can write its version of SAP Duet), and more.

Apparently when it crossed all the t’s and dotted all of the i’s on the agreement Microsoft made with a free-software-oriented organization called the Protocol Information Freedom Foundation in December 2007, it decided to just open the kimono and eliminate the middle man.
Can you say this means Microsoft is now open source? No, and Microsoft took time in the press conference to spell out its intellectual property rights (licenses will be reasonably available but not freely available). But anyone who argues about the differences at length (and many long-time Microbashers will of course) is strictly splitting hairs.

– Dennis Byron
Tags: Microsoft, open source, EU, Windows, Protocol Information Freedom Foundation

Red Hat needs to get red hot to make these numbers
Posted on February 14, 2008

Expanding on a plan it announced in November 2007 by which Red Hat (RHT) said it would capture 50% of the operating system market by 2015, the company announced on February 13 that it intends to also capture “50% of enterprise middleware workloads by 2015.” The goal comes as an interesting juxtaposition with the Alfresco Barometer Survey announced February 12 at the JBoss user conference. That survey indicated that JBoss was not even the most popular open source application server. If the latter is true Red Hat will have a tough row to hoe.

Let’s start with how Red Hat is defining middleware. The answer is very broadly (“more than just the application server alone”). To accomplish this goal using Red Hat’s broad definition of middleware, Red Hat would have to not only displace other potentially more popular open source middleware (e.g., Apache HTTP software) but billions of dollars of “closed-source” software currently in use across the world between now and then. (NOTE: The open-source vs. closed-source characterization is not meaningful in understanding or measuring the market as spelled out earlier in this blog post but the open source devotees still insist on drawing this distinction.)

To literally accomplish its goal, Red Hat would need to displace 50% of the $20-$30 billion worth of IBM (IBM) CICS and BEA (BEAS) TUXEDO shipped in the last 30 years, the $20-30 billion worth of MQ Series, TIBCO (TIBX), etc., shipped in the last 20 years, the $20-$30 billion worth of WebLogic, WebSphere and Oracle (ORCL) Application Server shipped in the last 15 years, the $10-$20 billion of BEA, IBM and so forth integration servers, development tools, ESBs, portals, and so forth. Add business intelligence software (if you use Oracle’s definition of middleware) and collaboration software (if you use IBM’s). Sorry but the aircraft carrier just does not turn that quickly.

As for how Red Hat is defining “50%,” the company says it means of that year’s “deployments of middleware technology.” Hopefully us reserachers will have some license-agnostic census software to measure that number by 2015 but to get there, do the math. No matter how you measure 50%, Red Hat would have to displace existing installed software at an astonishingly rapid rate over the next 6 years to reach its goal because the overall software market is only growing 5-6% (and the middleware market basically tracks the overall software market).

At the press conference announcing its intention, Red Hat said that its subscription model makes comparison with other middleware suppliers impossible. That is not accurate and part of the litany of open source movement claims that can be misleading to investors. Most of the current installed base of middleware has subscription maintenance revenue associated with it just like Red Hat’s.

In fact, at the press conference Red Hat showed an IDC estimate of $15 billion of middleware revenue in 2011; about 66% of that number represents subscription maintenance just like Red Hat’s (the other 33% in any given year represents license revenue for totally new business and add-ons/upgrades to the installed base). The equivalent IDC number for 2008 is probably in the $11 billion range, of which Red Hat’s total is less than $50 million.

(I am just guessing. The only access I have to IDC middleware market share data is via its press releases. On the other hand, it is a very educated guess since I popluated that database for many years and still do IDC backcasting and operating environment splits in my sleep.)

– Dennis Byron
Tags: Apache, open source, JBoss, middleware, MQ Series, Tuxedo

Slow and steady course for SOA on Wall St.
Posted on February 12, 2008

Leading bank and investment-firm CIOs, IT directors, staff technology gurus and the like spoke February 11 at the New York City conference “Web Services/SOA on Wall Street” Their theme could have been “SOA is not on Wall St. yet” but that’s probably good news for the supplier community that packed the conference with new technology value propositions. The slow uptake of SOA means the large systems suppliers such as IBM (IBM) and HP (HP) have not run away with all the services oriented architecture (SOA) business already.
Listening to the Wall St. information-technology (IT) gurus is important because they and their peers in The City and Switzerland and Singapore are usually two or more years ahead of the curve in terms of IT implementation. When it comes to investing in (or marketing) IT, finding out how it plays in Peoria first is not good advice. So in that light, where does SOA stand?

The general feedback from the presenters, most of whom are in varying stages of evaluating or implementing SOA for their firms, is that Wall St.’s own culture of silo operations is going to make SOA a tough sell. SOA helps eliminate IT silos but if the business itself is siloed, it does not want to share its IT resources with the other departments. The speed of migration to SOA by industry may be determined by the extent it is or wants to remain siloed for business rather than technology reasons.

Secondly, SOA does not mean web services to these guys. All of the buzz about mash-ups and social computing has a bad feel to it if an enterprise has concerns about governance, security, and risk aversion. Or has governments breathing down its necks with those concerns. That’s certainly true of Wall St. but increasingly, it’s just as true of Main St.

As is often the case, the vendors sponsoring the conference wanted to push the IT users on to the next set of buzzwords. To heck with the title, “Web Services/SOA on Wall Street,” how about complex event processing and Enterprise 2.0? Wall St.’s answer, “let us walk before we run.” One of the best lines of the day: Look at the complexity of the Google (GOOG) infrastructure. And that’s “just simple event processing.”
The implications of this, as I have written about here and here and elsewhere, is that SOA is going to permeate the world’s computing infrastructure slowly, just as client/server computing did beginning 20 years ago but probably not reaching a tipping point until after the Y2K scare passed. Similarly, by the time SOA happens, the industry will be on to a whole new set of buzzwords. SOA will eventually have a major technological impact but very little impact on investment strategy.

– Dennis Byron
Tags: service oriented architecture, SOA, Enterprise 2.0, mashup, CEP, web services

Middleware indie/open source Iona says its biting the dust
Posted on February 8, 2008

UPDATE: Or maybe it’s Novell? (NOVL)”

Iona (IONA) announced February 8 that it is likely maybe possibly thinking it might get acquired.

It’s a two-fer winner for this analyst’s recent predictions in that the likely acquisition is

  • An example of another independent middleware player biting the dust after the Oracle-BEA acquisition and right on the heels of Workday’s acquisition of Cape Clear on February 5
  • A likely example of an open source software (OSS) supplier pairing up with a proprietary supplier

I do admit that the first prediction–the rapid reduction of independent middleware choices for users–is not that earthshaking. It’s the natural result of the commodization and coagulation of the middleware stack that I began researching at IDC in 2000. Also I appreciate the comments on my earlier blog post by Giva Perry of gigaspaces who points out that there is a new generation of middleware emerging that could change the market dynamics. I will dig deeper into Giva’s ideas in upcoming research.

The second prediction vis a vis Iona is a little more out there. There’s nothing in the press release saying the possible suitor is a “proprietary” software supplier or even a software supplier. Apparently under Irish or EU law, we’ll know pretty soon who the suitor is. It could be Red Hat (RHAT) in which case my prediction is wrong. More likely it’s IBM (IBM), TIBCO (TIBX) or Sun (JAVA). Most likely, it’s a telecom (where Iona has a big presence) or a packaged applications suppliers, acquiring Iona for the same reason Workday acquired Cape Clear.

Of course there’s another disconnect: Although Iona has positioned itself as an OSS player for the last few years, it is clearly in the camp we describe in our research as proprietary/OSS hybrid (see summary feature articles that we release via ebizq.net). So when the OSS blogosphere goes crazy congratulating itself after the likely maybe possible deal goes through, don’t ratchet up your OSS valuation charts. Iona has been in the middleware business since even before the term middleware was widely used and has only adopted an OSS facade in the last few years.

– Dennis Byron
Tags: Iona, open source, TIBCO, gigaspaces, OSS, IBM

Sun speaks SOA for feds
Posted on February 7, 2008

Sun (JAVA) ponied up some real money this week to join the Object Management Group (OMG)-administered SOA (service oriented architecture) Consortium. Other sponsors include Cisco (CSCO), IBM (IBM), SAP (SAP) and Sparx Systems. I believe BEA (BEAS) was a sponsor when the consortium began in March 2007 but it does not seem to be listed any longer.

Sun’s press release raised two questions in my mind:
The SOA Consortium is the group that said at its founding that it was going to go out of business in 2010 (“a closed-end operation” was the term the OMG used). So why did Sun join now?
Who to hell is Sparx Systems and what are they doing in the pool with those sharks?
The SOA Consortium says it is all about “Redirecting the industry conversation to business-driven SOA.” I think the SOA Consortium—perhaps because of IBM’s apparent instigating sponsorship—is trying to overcome the common technical use of the words “services” and “architecture,” trying to position SOA as a business strategy as part of an IBM marketing messaging tactic.

The better analogy in my opinion is to compare two eras:
In 1995: client/server was the architecture behind a business process re-engineering (where have you gone, Michael Hammer?) business strategy

In 2010: service oriented architecture is the architecture behind a business process management (BPM) business strategy

BPR was all about using a single large monolithic package such as SAP R/3 in an enterprise. BPM is all about tying together multiple packages within and among enterprises.

As part of its kick off in 2007, the SOA Consortium interviewed CIO and CTO focus groups. I interpreted the consortium’s findings in the BPR vs. BPM context to mean that SOA would cause a fundamental market change for packaged application providers. In the client/server generation, suppliers gave away the platform (e.g., ABAP, PeopleTools) to sell the application modules (R/3, PeopleSoft HR). Users say they expect application developers to do the opposite in the SOA generation: give them the services (SAP ESA) if they acquire the SOA platform (NetWeaver).

Perhaps that’s why startup ERP SaaS provider Workday acquired the Ireland-based Cape Clear middleware maker this week. (Oh by the way, scratch another independent middleware player off this recent list.)
The answer to why Sun is getting involved with this evolving SOA buzzword game seems to lie in the Sun executive quoted in the press release. It didn’t come out of the Sun front office but from Bill Vass, president and COO of Sun Microsystems Federal. So there’s probably a tactical U.S. government opportunity somewhere in there with the pony.

As for the second question, according to an excerpt of an IDC report released as a promotional piece by Sparx in May 2007, it is a private 10-year-old Australian software development tool maker that is not even that well known in Australia. Its tools are mostly deployed into midsized development organizations according to the IDC article.

Sparx’s product is based around Model Driven Development and the Unified Modeling Language and is platform agnostic. That is, it works with the Eclipse/Java-oriented guys that Sparx is in the shark tank with at the OMG, as well as with Microsoft (MSFT) Visual Studio and Visio.

Interestingly, I can’t find the term SOA anywhere in the IDC report on Sparx? Go figure.

– Dennis Byron
Tags: Microsoft, SOA Consortium, Sparx, Object Management Group, UML, SOA

NYC financial analysts just “can’t handle the truth”
Posted on February 4, 2008

Steve Ballmer, Microsoft (MSFT) CEO, and Chris Liddell, CFO, presented to financial analysts in New York City on Monday, February 4, 2008 as part of a long-scheduled event.

It would have been kind of hard to top Friday’s news about Microsoft’s unsolicited bid for Yahoo (YHOO). And they didn’t. But they did fill in some of the financial blanks on their plans for Yahoo two to three years out and they explained dollar for dollar plans for Microsoft’s future. To be discussed, the investment potential had to top $750 million in future potential contribution margin.

Ballmer’s and Liddell’s claim is that Microsoft is the only technology company currently with the assets to mash consumer and enterprise computing together. I agree with the underline on “currently.” They also say buying Yahoo is just a means of getting them to their objective of profitable scale and capacity faster. I agree and believe it to be the best indicator yet that Microsoft wants to move rapidly to services provision from technology provision (otherwise the Yahoo acquisition makes no sense). This move from technology to services is the opinion reflected in the current Research 2.0 Microsoft report, released last month.
One problem is that Microsoft still calls itself a technology company rather than a services company. I think that’s because Ballmer and Liddell believe that the NYC financial analysts want to hear Microsoft described that way. Like Tom Cruise in A Few Good Men, they just “can’t handle the truth.”

At least I hope that’s the case because unless the Microsofties say it out loud a few times, a services mentality is hard to internalize.

A few other tidbits:
I think they succeeded in conducting the whole hour-long presentation with Q&A without saying the word “Google” (GOOG)
Look for very aggressive pricing for the virtualization functionality within Server 2008. Ballmer is looking at the pricing as an investment in buying market share.
Microsoft thinks it took market share away from Linux in 2007. Actually Research 2.0 studies discussed here say both Microsoft and Red Hat took share away from Unix suppliers (meaning the old-line systems suppliers)

The Software plus Services strategy is being renamed the Microsoft “cloud computing” strategy. Same strategy but new buzzword.

– Dennis Byron