Friday, January 23, 2015

IBM - a Once and Future Giant or Is It a Has Been?

For a company founded in 1911, you would think that 2015 would be a bit late for a mid-life crisis.  That, however, is what appears to be happening with IBM (NYSE: IBM) since they no longer appear to know what they want to be when they grow up. That's assuming, of course, that IBM is able to recover from some extremely poor strategic decisions, coupled with an inability to deliver what the market truly requires.

Consider that, since the 1960s, IBM was the computer hardware company, driving the business technology world through the information revolution and into a world where computing technology is a given in our everyday lives.  The phrase, "Nobody ever got fired for buying IBM" was inevitably heard in any management discussion regarding a decision on acquiring new technologies.  Computer hardware and software was IBM's domain, and it was IBM's to lose.  And lose it, they did.

After driving their competition out of business, IBM emerged as the sole provider of Mainframe hardware (including the current z/12 mainframe family and the just-announced z13 mainframe family.)  They are also the sole provider of the mainframe operating systems as well as the only mainframe transaction processors (CICS, IMS, or Websphere for z/OS.)  There's only one problem.  IBM is not increasing their mainframe footprint, and existing mainframe customers are gradually migrating new applications into the non-mainframe space.  The mainframe market is shrinking rapidly, and you can thank IBM's own internal marketing competition for that shrinkage.

IBM managed, though the early part of the 21st century, to undercut and cannibalize their own advantage.  They introduced the P/Series hardware technology as a direct competitor, encouraging mainframe customers to migrate to the faster (at the time) mid-range technologies running their AIX operating system.  The amount of Level 2 cache on the the P/Series boxes was a fraction of what was on the mainframe, though, so multi-tasking was not a forte in the mid-range space.  Websphere Application Server (WAS) is the transaction processor of choice off the mainframe, however it's a major resource hog and can't compete with CICS when it comes to scalability, performance, or stability.  (Ask any WAS specialist how often they hear "A JVM is looping."  Stability and WAS are mutually exclusive.)  Yet, the number of WAS licenses dwarfs the number of CICS licenses, thanks to IBM's efforts to increase P/Series and x86 based sales.

The industry trend now, is to migrate off the mid-range P/Series that IBM sold by the bushels a decade ago.  AIX as an operating system was deemed too expensive and is rapidly losing to Linux as the Unix-based operating system of choice.  Linux can run on all platforms - including the mainframe - and the application code running under Linux is highly portable.  This is where scalability is factoring into the demise of IBM.

Unlike the monopoly IBM was able to develop in the mainframe space, the non-mainframe server space is highly competitive.  HP is emerging as the market leader in that space.  IBM was #2, but in a move that still mystifies me, IBM sold that advantage to Lenovo.  The x86 server space is growing rapidly, and IBM gave it away.  Dell is also in the top 3 in that space, giving customers plenty of options for the lower cost but highly scalable non-mainframe servers.  When you look at the hardware filling all of those expanding data centers, you'll find endless racks of x86 based servers.  Why would a major hardware company even think of abandoning that huge growth market?

In the database space, there are two main competitors and a bunch of also-rans.  You're looking at either Oracle or DB2.  On the mainframe, running under z/OS, DB2 is the only option (although there are still plenty of mainframe shops storing data primarily in VSAM, which comes with the operating system.)  In the non-mainframe space, however, Oracle is a huge factor.  In 2014, the two were neck-and-neck for market share with a very slight edge to IBM's DB2.  Oracle, however, is gaining rapidly.  With some of the acquisitions Oracle - the company, not the database - made over the past few years, they are now a force to be reckoned with in both the hardware and software spaces.

So what are IBM's plans for the future?  Well, we know they've abandoned the chip market and they've abandoned the x86 market.  They are hanging their hat on remaking themselves as a "Cloud Services" provider, a Software company, and, of course, they are touting the just-announced z13.  It's not enough.

To be a viable cloud provider, they should never have abandoned their x86 business.  The competition in that space is huge, and by selling the x86 to Lenovo, they have relinquished a huge cost advantage in that space.  Let an HP or a Dell partner with either Oracle or EMC, and you'll have a stake driven through IBM's cloud-based heart.  IBM's Global Services division - their technology outsourcing business - is hemorrhaging market share to the India based outsourcers.  Let an Infosys, or a Wipro, or a TCS partner with one of those combinations I mention above and IBM's cloud dreams are down for the count.

On the z13 front, IBM mentioned all of the major buzz-words that make senior executives salivate.  They mentioned "Cloud" and "Mobile" in the same breath, and that always makes the non-technical executives swoon.  Under the covers, though, you merely have the next release in their standard mainframe offering with some improved performance, improved throughput in their FICON channels and their zIIP specialty engines.  They've made some nice improvements in the zBX blade-bolt on to the mainframe.  But, at the end of the day, it's still a mainframe, and you're not going to see anyone flock to the z13 to satisfy a cloud or mobile requirement.  For existing mainframe customers, it may delay a migration to the distributed world, but it's not likely to attract new mainframe customers, and as companies world-wide continue to squeeze their expense budgets, new customers is precisely what IBM requires.  They're not able to squeeze more revenue from their existing customer base. 

The direction IBM is taking at the moment does not lead me to believe they are on the right course.  Rather, I see them heading in the exact opposite direction from where they need to go.  For now, the only thing IBM has going for it is its name.  Of course, there was a time when one might have said the same for Sperry Rand.

No comments :