Comments on CIO's Guide to Cloud Computing and On-Demand | Appirio: Cloud Computing: Hummer or Prius?

Great dialogue, Ross & Jay-- here are some tho...

2008-08-12T22:59:00.000-07:00

Great dialogue, Ross & Jay-- here are some thoughts:

- Ross, I don't think it makes sense to invest in building and owning non-core infrastructure regardless of your size (or funding situation).... even the biggest companies use electricity off the power grid, even if they have the resources to build their own power plant. It's Moore's core/context argument-- focus on what is going to make you distinctive, outsource everything else that you can. Most companies (even technology companies), aren't going to compete on their infrastructure. Jay's right-- let someone else worry about that.

- Jay, CIO's are certainly concerned about switching costs. But after digging into the survey, I think the answer is even more simple: CIO's correctly view cloud computing as a means, not an end. Of course CIO's prioritize their business objectives (e.g., cost cutting, app integration, BI, CRM, etc) over the underlying technology required to achieve those objectives. That's all the survey reflects.

To extend the car analogy: Would you say that you are more interested in getting to work quickly and economically, or in deploying cutting edge hybrid engine technologies?

In getting to work, of course-- but if you're smart you'll still prefer a Prius to a Hummer!

I'm not sure if IaaS is necessarily the always...

2008-08-12T22:36:00.000-07:00

I'm not sure if IaaS is necessarily the always the better choice either - even with a company that's more "grown up". (admittedly, this is also a broad sweeping generalization).

I'm picturing the non-mission-critical small department in a huge conglomerate whose IT requests gets buried under the other high-priority urgent requests from larger departments. They're almost like a tiny startup living in the belly of a larger beast. I think the pay-as-you-go model makes it attractive to this segment, since it doesn't require a large upfront budget. They can just "test the waters" and see if they really want it.

On the topic of getting into the nuts and bolts of hardware - Google's server farm needs no introduction. They are not specialized embedded servers with special Google chips - it's all commodity hardware, your standard-issue Intel x86, essentially a Wintel box (sans the "Win"). And it works well for them, an understatement since their server farm *is* one of their competitive advantages. They did a research on hard disk failure rates.

http://usenix.org/events/fast07/tech/pinheiro.html

http://lwn.net/Articles/237924/

"The annualized failure rate was much higher than estimated, between 1.7% and 8.6%"

So if you have at least 100 hard disks in the in-house IaaS, be prepared for at least 1 or 2 (up to 8 or 9 in you're unlucky!) HDs to act funny every now and then. Other things to worry about: cooling, power, security, physical security, fire, .. personally, I don't want to worry about that, and rather let the infrastructure experts deal with that (it's their core competency after all, mine is somewhere else and I rather focus on that).

Another article worthy of mention here -

Facility costs over overlooked in IT investment process: http://www.forbes.com/2008/08/10/cio-cheap-servers-tech-cio-cx_kb_0811servers.html

A blurb about TCO: "Spending $2,500 on a server really means spending between $8,300 and $15,400 in facility capital to provide the necessary space for housing the server and powering it."

All in all, what comes to my mind right now is something Jeff Bezos once said to the audience of Startup School at Stanford U, consisting of young twenty-something web entrepreneurs, when he was asked what he would do about competitors.

To paraphrase, (since I don't remember the exact words) he envisions that this space will have "multiple winners" -- in contrast to a one-Microsoft-way dominating giant that takes it all. IaaS will have it's place in the value chain, as does cloud computing, SaaS, .. etc.

###

Ryan, do you think that the problem might not so much be that CIOs don't know what it is, as it is that the burden of letting go of the comfy on-premise solution everyone in the company has grown so accustomed to over the years? That is to say, the one time total cost of switching may be perceived to exceed the potential savings?

Cloud Computing may not be the answer to all CIO&#...

2008-08-02T11:18:00.000-07:00

Cloud Computing may not be the answer to all CIO's problems! Cloud Computiung is often better suited to the startup rather than the enterprise.

In the past if you came up with the next ‘killer app’ you would be faced with a serious decision…do you try to build your business on a shoe string and risk the chance of being a victim of the Slash Dot effect or do you try to navigate the slippery road of equity investment funding. In our case we decided to break the mound and use IaaS to give us infrastructure and scalability in a cash flow friendly
way.

We have been doing this very successfully and we have eventually bootstrapped ourselves into profit. My recent calculations have shown me that once you are actually taking money the economics of IaaS can change significantly. While having unlimited scalability in a cash flow friendly manner is advantages in the early days it can often be cheaper in the longer term to transition the infrastructure in-house (or to a co-location facility). While this calculation is dependent upon how you application uses hardware it is generally true that if
you need RAM it is cheaper to buy it by the stick than by the hour.
So, I have started to look at buying our own infrastructure
transitioning our application to this. The ultimate goal is to move
the bulk of the application to your own infrastructure and still use
IaaS for any additional or burst capacity.

So where previously the startup had the ‘shoe string’ or ‘VC’ route
they now have an additional option…the ‘bootstrap and transition’
route.
More here:

http://www.spoutingshite.com/2008/08/02/iaas-the-answer-to-an-internet-startup%e2%80%99s-dreams-but-not-a-co-lo-killer/