The Industrialization of IT

I sat on an industry panel for a Goldman Sachs investor conference today focused on next gen IT and cloud.

Although we were supposed to be covering investment themes in the hardware sector as the industry industry transitions, I ended up (as usual) probably going too far afield.

Now, as I reflect on the session, I realize that were likely some important macro themes worth sharing and discussing.  See what you think?

Trouble From The Outset

As with any panel, it usually starts with an ostensibly easy question that sets the stage for everyone in the room, i.e. what’s going on?

I keep coming back to Simon Wardley’s comparison of cloud computing to the industrial revolution.  There’s more to that analogy than meets the eye.  Fits nicely with Nicholas Carr’s views in the Big Switch.

We’ve begun to witness the industrialization of IT.

Simply put — this is no minor industry transition.  How we build, operate and consume IT will likely be forever changed in a handful of years.  Everyone will be affected.

As with any industry transition, there are winners and losers.  And, of course, given the audience, they were intensely interested in investment themes to consider.

In-The-Stack vs. Best-Of-Breed

I was on stage with two other storage vendors, and I wasted no time jumping into the deep end with the controversial observation that the industry was quickly consolidating into “stacks” — legacy ones such as HP and IBM, newer ones such as VCE as well the potential for an Oracle/Sun stack.

As I write this, the news on Twitter is that HP is buying 3Com for $2.7 billion in cash.  The drumbeat of consolidation and stack-building continues.  The real surprise is — why is anyone surprised?

I offered that technology players in a popular stack would have an inherent business model advantage in terms of scope, scale, integration and efficiency.  As the industry consolidated, more of the spend would move to stacks, and less to traditional best-of-breed piece parts vendors that supplied one subsystem or the other.

The two storage vendors on stage offered up a vigorous rebuttal (all part of the fun) by saying that the only way for IT organizations, system integrators and service providers could differentiate themselves going forward was  by selecting and integrating the best of breed technology, ostensibly with their product.

I countered by observing that any best-of-breed player — no matter how good — faced structural challenges in their business model in the very near future — going to market costs, convincing customers to break up their chosen stacks, costs associated with integrating and supporting all the major stacks, getting a fair return on their R+D investments, and so on.

At some point, it becomes more about the best power plant (to use Nick Carr’s analogy) and less about who makes the best rotor, bearings, etc.

It’s All About Service Providers

Very quickly, the discussion turned to service providers and their new role going forward.  I described them as a new and important “channel” of distribution between companies that build technology, and organizations that will consume it.

There will be some service providers who carve out an area of vertical or application expertise as their competitive advantage (Salesforce.com comes to mind).  These outfits will want many of the same things that any large-scale IT organization wants.

More interesting is the newer wave of “horizontal” service providers: telcos, outsourcers, etc. who are very good at standing up IT infrastructure at scale, and running it extremely efficiently.

However, very few of these newer large-scale players have a way to reach IT organizations effectively.  Most enterprise IT organizations want a “high touch” relationship with their vendors, especially when considering something relatively new.

For a service provider that’s focused on being cost-advantaged, that’s a business model problem, since building such an organization is an expensive proposition.  Better to sell your cost-advantaged service through someone else who already has a customer relationship, e.g. a big enterprise IT vendor or their partners.

Put it all together, I said that the likely winners in this next phase of IT would be the players who could put together not only a compelling stack, but a compelling go-to-market ecosystem to take it to market.

And Then It Got Back To Storage

No surprise, EMC always gets put on panels with other storage vendors.  So, sooner or later, I knew we’d be talking about storage.  The other two panelists were very focused on storage implications of the new cloud models.

Both did an admirable job of driving their talking points home — how well positioned their unique technology was, how many service providers were using their products, and so on.

Always the controversial one, I suggested that we were in the very early phases of this industry transition, and to claim leadership or victory was a bit like Netscape declaring that they had won the browser wars many years ago.

Given that one fellow panelist represented a file-system-based storage product (NetApp), and the other fellow panelist represented a block storage product (3Par), I went for the “none of the above” ground by suggesting that object-based storage might be the winner at the end of the day.

Which got us into a very interesting discussion indeed.

Consider The Next Wave Of Enterprise Applications

Given that I probably hadn’t stirred up enough controversy, I started to bring out some heavy ammunition.

Consider, for example, that an enterprise application created with modern tools can ideally be thought of as potentially “cloud ready”.  These applications may not think in terms of blocks, files, or even databases.

Indeed, much of the legacy storage, operating system and database technology being fiercely debated today may not be all that relevant after 3 to 5 years of application rewrites in these newer frameworks.

That gave everyone some food for thought

Questions From The Audience

One expected question was around concentration of buyers and resultant loss of pricing power.  As the theory goes, if there are fewer large-scale IT buyers going forward, doesn’t this mean that IT vendors lose pricing power?

As if we as vendors had any to begin with …

I presented an alternative scenario — selling and supporting a smaller number of larger consumers of IT is inherently more efficient from a business model perspective.  And any vendor with a considerable R+D budget now has a clear use case target to create relevant differentiation that tends to get magnified at scale.

Someone else tagged into the ever-perennial “isn’t IT getting commoditized?” question.  I always like to take a crack at this one.

First, if you look at the last 40 years of IT, the cost-per-unit of IT has come down spectacularly.  And, at the same time, our economy ends up spending more and more on IT every year — economic hiccups excluded.  Why would that change in the future?

Second, if you agree the premise that we’re seeing the industrialization of IT, there are plenty of opportunities for smart players to capitalize on the shift in a big way.

And Then Our Time Was Up

I don’t know whether the discussion was what the organizers were expecting or not.  We went far outside our brief in terms of topics discussed.

Sorry about that, folks.

I’ll be interested to find out whether or not the audience thought it was a good discussion, or not.