|
|
The End Of The Storage Business?
By Steve Duplessie
Expert Author
Article Date: 2007-10-17
Where to begin? It's been a long 10 days or so, so this rant may be a tad longer than most. First, I ran off to Aruba with my lovely wife and two little ones for a week.
Great place, Aruba. One happy island. I had every intention of spending the week under a palm tree with my laptop getting a lot of work done, or actually writing the cancer book I've been promising for years. I did neither. I did spend a lot of time soaking up sun, cocktails, and pondering IT issues. Here's what I came up with:
The storage business as we know it, is over. I don't mean you won't continue to buy billions of dollars of storage products and services, but we're at one of those pendulum swinging crossroads as an autonomous business. Who got me thinking this way? Sun, of all companies. Only Sun could say something in such a wrong way, for the wrong reasons, to the wrong people, but end up being right without necessarily knowing why. Right before I left for 'de islands mon' Sun, in typical chaotic fashion, insisted on getting on my calendar for a huge, important, giant announcement they were going to make - "tomorrow, it has to be tomorrow - and no, it has to be Steve, and no, we can't tell you anything about what it might be related too....". Somehow, our client relations folks, with the patience of saints, were able to move things around so I could get on the phone for 30 minutes, from my car - which is a very loud convertible, to hear the latest revelation from the west. Their announcement, also known as Jonathan's blog, was to be around the fact that Sun is rolling its storage business back under its systems business. Whoopdie doo.
So forgetting the fact that it really wasn't worthy of an announcement at all, nor did anyone pick it up, nor did anyone even ask me about it (that is one company that is in serious need of re-inventing how it goes about messaging. They make HP marketing of the 90's seem downright brilliant sometimes), it did make me think about some things at the swim up bar.
20 years ago there was no storage business. It was a systems business that used storage as part of the overall platform to execute some applications. Companies competed on what their systems could offer for application functionality, typically embedded. The term "peripheral" was used to represent add-ons such as incremental storage capacity, but no one really bought storage independently from the system. Sun changed the "systems" business - namely it suddenly offered faster, cheaper, better new ways for engineers (mostly) to do their jobs than traditional business systems - minicomputers - offered. Unix became cool. Sun changed the game - they got the whole distributed computing thing moving. Right or wrong isn't relevant - it was entirely "blue ocean". They didn't focus on "my processor is better than yours, or my O.S. is better than theirs" - until later, and of course, that's when they went amiss. During the game changing era of Sun and distributed computing, people bought storage for Sun workstations from Sun for the most part - because that's how things were done. Peripherals.
So Sun completely changed the face of the commercial computing game - a well established gazillion dollar a year business with lots of very big companies in it. At the same time, other external factors were creating yet another game changing market opportunity - to change the way people thought (and bought) about storage. EMC got people to stop looking at storage as a peripheral check box when you bought your next mainframe or minicomputer (which is still one of the most remarkable feats ever in the history of infrastructure if you ask me). Two, NetApp looked at Sun's success, and figured out a way to segregate the decision on file server storage. They solved totally different problems (one for monolithic core computing, the other for the new wave of distributed networked computing), in totally different ways, on totally different coasts - and both have done fairly well financially because of it. You could argue that in the purest sense, EMC has dominated the block core world ever since, and Netapp the file serving world. Two halves to a market that didn't even exist at the time, which have yielded both companies outrageous fortune and spawned a huge industry. Blue oceans. Both did even though both were a piss-ant size wise compared to the incumbent systems guys they were stealing from, and both had to not only sell against those incumbents, but they had to teach the market an entirely new way to buy. That's hard stuff to do.
For the next twenty years, the "storage" business has been a 20-50 billion dollar annual love fest - tracked as a totally independent and autonomous market.
The storage "business" happened because the technology focus and specialization required to optimize that piece of the system function - I/O - had specific problems that were not able to be generalized any longer. It isn't unique, it happens all the time - we live in cycles moving from generalist technologies to functionally specific ones. Routing functions ran in the operating system, until Cisco ripped it out and created a function specific gizmo. Data center caliber storage went through the same cycle. For the same reason no one ever has taken Cisco down in routers, no one has taken EMC down in the core data center nor Netapp in the file server market, nor IBM in the mainframe space, nor Oracle, and on and on. All have competition - each having "better" stuff. It doesn't matter. Those games are over - so without a new game - a blue ocean, things will largely stay as they are.
So where was I? Right, no more storage business. EMC and Netapp, or Compellent, Lefthand, etc. etc. are not storage companies. They are systems companies. They run specific functionality on processors to execute applications - that just so happen to be geared to storing, moving, protecting data that will be used by some other processor running some other functionality for some other area of interest. They build computers - almost identically to those back in the day. Storage guys used to write their own RAID code, because that was what made them different. Prime and Wang and DG and DEC and HP and blah, blah, blah used to all make their own processors and operating systems, because that's what made them different. Nobody buys one guys RAID 5 box vs. another guys because of the RAID code anymore. 99% of the processors used in commercial computing are made by 3% for the guys who used to have their own. Who in their right mind would build a general purpose operating system today? No one. Those games are over.
Sometimes the game is changed because of whole new way of doing something. Sometimes it's about technological innovation (Riverbed). Other times its a simple matter of applying technologies in a whole new way (VMware), and other times it's simply kicking some slow, fat, dumb butt out in the market (Add your favorite example here). When you have all three, you tend to make a pile of dough. Pure luck isn't sustainable - you have to be able to see the forest through the trees, and you have adapt to the world around you.
Netapp is a neo-classical systems company. They use commodity components and put their money into their IP, which is their operating system and the applications that run on it. Those applications happen to be storage applications but what's the difference between them an any other systems company - other than the fact that they don't let anyone else write applications directly to their O.S.? EMC is the new version of IBM or HP - they not only are a systems company, they are software company, service provider, and so on. Larry Ellison, in his recent acquisition targeting BEA, said "people want to buy software from one place" - and he's right. The same holds true for all the other stuff as well.
So of course Sun should roll up storage under systems. They might not know how to say things properly, but they have piles of talent and a market built around their processor and their operating system. Forget whether or not it can or will work for them, but the theory is exactly right. Sun, IBM, EMC, etc. need to think of themselves as systems companies who delineate their activities by the functionality provided to the individual markets they serve, not as a storage, systems, software, etc. companies. Those who prosper in the next frontier will be those who understand that we are moving on - it ain't about the micro function or gizmo, it's about the macro. People simply can't continue to live in the weeds in IT, they need holistic systemic infrastructure to ever hope to bring real value to their organizations. System companies are going to be what's required to puppet master all the micro gizmo function. You have to be able to give the people what they want, whether you think it's right for them or not.
Wyeth-Ayerst, the giant pharmaceutical company, makes birth control pills and baby formula. How smart is that?
For those who don't know, the Blue Ocean Strategy is a book I'm fond of, regardless of industry.
Comments
About the Author:
Steve Duplessie is the author of the "Steve's IT Rants" blog, and the founder and Sr. Analyst of the Enterprise Strategy Group.
|
|
