Public vs. private is the hottest topics going in the cloud blogosphere. The reason is simple: customization is the biggest market in I.T—and keeping things somewhat custom is in a lot of peoples interest.
In my time as a corporate development monkey at an infrastructure vendor, one of my best ‘ah ha!’ slide decks asked the audience to stack rank the major categories of IT spending, with this slide as the revealed answer next:
Yes customizing integrating and maintaining IT costs tremendously more than the hardware and packaged software. These numbers reflect only outside firm spend, so double or triple this to see the full impact when including internal headcount. You quickly arrive at an obvious fact: I.T. is far more a human process than a technology one.
This perspective has stuck with me since I created the deck in 2007. Its why I call much of the code automated cloud computing phenomena a collision with traditional vendors and IT. If a more automated and standardized approach is really going to sweep IT in the next two years it would disrupt hundreds of billions of dollars in spending and purchasing habits—and the incumbents, both from a human and vendor perspective aren’t going quietly.
The real debate: how much of the customization is wasteful?
The best argument for private cloud development is a quick glance at the status quo. A market willing to spend 400B++ a year customizing, won’t suddenly conclude a bare bones public cloud is a cheap and cheerful alternative. So yes, to all of the private cloud proponents, you are right, they will have a huge role in the future of IT spending. The overall net cost of applications today is simply too massive to make the potential risks of simultaneously migrating to a new architecture and an outside provider worth it.
But, although the change may occur over geologic time with the status quo being favored–new application development and consolidation are the wind/water/tectonic metronome of geologic time in IT. So to accurately articulate the coming impact and segmentation of the cloud we should carefully study both new apps and consolidation trends.
My Angle: Its really a debate of consolidation vs. scaling
Last week @roidude detailed a 100% vitualization project ROI study this week in a blog. It makes a very compelling argument for consolidating non resource intensive applications to virtual servers, in fact his project promises a savings of $2.1M out of $2.3M or an ROI (dude!) of over 500%.
Whoever is in charge of this IT consolidation project would have a lot to brag about—but whoever is managing their application portfolio will get a LOT of questions from me. The only reason its possible to consolidate this many applications onto so few servers and get such a galaxy beating ROI% is because the demand on the applications is not growing.
The era of “hey lets build an app for that” vs. the era of “there is an app for that”
With the emergence of meaningful SaaS portfolios the ROI of investing in creating a custom application and supporting it vs. using a world-scaled public cloud based SaaS application is dwindling. This is a big change from the old days when many of today’s process applications were created.
The quantity and demand profile of these legacy applications create a huge consolidation opportunity—but I’m not ready to declare a cloud consolidation architecture corresponds to the right environment for unique and high growth applications. They are different animals and I don’t hear this discussed nearly enough.
So the first thing we should ask when someone begins the private vs. public, open-source vs. vCloud can of worms is:
Q:How fast is your application growing? What is its potential scale?
@randybias expertly detailed a bifurcation of public vs private clouds in terms of feature sets, but I didn’t see any color in terms of application scale requirements:
There will be two paths for clouds: premium & commodity. Premium clouds will focus on the enterprise and delivering value they are concerned about. Commodity clouds will largely be forced to compete on pricing and features irrelevant to the enterprise. VMware’s vCloud will be the dominant player behind the firewall because there is no credible contender.
My cloud Rorschachtest is different. I see a bifurcation between fast growing, world scaling, applications and proprietary lower & scale/growth internal applications. Google, Facebook, WordPress all teach the religion of building solid parallel programming primitives right into the applications. They spend their customization dollars making the applications cloud natives. Greg Papadopoulos called this trend years ago:
The aha! moment came when Papadopoulos realized that there are really two different application sets driving computing demand: one consisting mostly of newer Web-facing applications driving exponential growth in both user demand and computing requirements; the other comprising back-end systems that are growing at more historical rates. “All this is really about which side of Moore’s Law you’re on,” Papadopoulos says. “If your applications are growing faster than Moore’s Law, you’ve got a fundamental set of issues about scale and power. If they’re growing slower than Moore’s Law, you’ve got all kinds of opportunities around consolidation.”
If you are building a cloud to consolidate your slower growing core then the ‘feature’ you need most from a cloud is application portability and the ability to virtualize without any change in the application. Its likely you spent eight figures getting some of those applications ‘right’ and vCloud is probably exactly the technology you need to float them downstream towards the cloud.
I challenge the industry to dream bigger than consolidation. As @lewmoorman wrote this week, all companies are becoming software companies as the consumer cloud explodes. This brings to mind the second geologic force, new application development. This is the area where private clouds and the vCloud product are most vulnerable. Being dominant in consolidation doesn’t simultaneously win new application development—and here is where massive ‘commodity’ clouds can really shine. When the application is built with a cloud in mind and can automatically load balance and spin up resources programmatically (ex:Hadoop) it removes much of the feature burden on the virtual machine. These world scale, cloud native apps are the interesting part of the market.
If you aren’t busy writing the world scaling app for your industry somebody else soon will be. Ask restaurant owners who now pay the commissions funding Open Table’s IPO. Somebody in the next two years will bring commodity cloud economics to your industry. (Why not you?)
Private public Red Herrings
If you want to build a massive volume standards based cloud in your own data-center that’s great—I don’t think location is really worth arguing over—but standards and ecosystem are. Its the participation in a larger ecosystem of interfaces, tools, best practices, and software that matters. But do you really have the process and governance discipline to maintain standards and avoid someone trying to prompt a promotion for their uniquely better way of rolling out your cloud? Do you have the organizational discipline to hold back the #meatcloud at the gates? Are these risks greater than those inherent in a public cloud?
Instead of debating location debate ecosystem.
While I can’t refute or answer every potential public cloud concern raised in the discussion, I can point to the $200 billion dollar application customization force in the industry as a potential solution. If customization/integration is the primary spend in IT, why not use it to help industrialize what is still largely a hand crafted process? So we have a data portability problem? Security and trust issues? Trust and regulation?
Why? Because I don’t believe the old way of doing things is designed to power the next generation of world-scale application processes. The alternative is a series of sustaining innovations around consolidation—and while these are economically vital—I don’t believe they can drive the next wave of growth in IT. Tom Siebel had an interesting quote in the NYT:
No new technological advances, he believes, would impel I.T. customers to replace the computer technology they already had: “I would suggest to you that most of what’s going on today is not very exciting.”
I find the ability to build an application with world-scale built in pretty exciting—but he is right. Replacing what we have isn’t the exciting part, using the virtues and open communities building up around public clouds and their architectures to build what’s next is.