Friday, April 17, 2009

McKinsey Report Misses The Mark on Cloud Computing

Couple of McKinsey guys have kicked up a micro storm in the Cloud Computing world with a presentation that claims that moving your data center into the cloud is not advantageous for IT. I am quite confident that others at McKinsey would differ - perhaps those presenting at another conference that is not catering to a server-hugging data center audience.

As TechCrunch summarizes - 'The report paints cloud computing as over-hyped and maintains that cloud computing services like Amazon Web Services (AWS) overcharge large companies for a service the companies could do better on their own. The study also says that while cloud computing is optimal for small and medium-sized businesses, large companies will spend less if using traditional data centers. Virtualization is the optimal way to go, says McKinsey, and by implementing virtualization in-house, corporations can reduce costs when factoring in depreciation and tax write-offs. Virtualization, which McKinsey says can boost server utilization to 18% from 10%, lets you treat one machine like many, by carving the servers into many virtual engines, so that software can maximize power from one machine and add scalability. Not only is this cost-effective for companies, but cloud computing takes advantage of virtualization.'

I believe that this report misses the whole point when it comes to Cloud Computing.

What McKinsey duo did was assume CIOs take the same exact crap software and run it on exact same excessive number of nodes provisioned for the full month at peak capacity numbers - and move that to the Cloud. That's just wrong!

This is even worse than ASP model where at least the vendors provide 'managed services' and even then it was a big failure. The real value of cloud computing kicks in when you leverage a full application platform (such as to either buy applications written to take advantage of a multi-tenant platform or write your applications on this new stack.

Why McKinsey Report is Flawed?

There are some key things that the McKinsey duo ignored:

1. Elasticity vs Provisioning for Peak Load: They assumed that if you buy 50 servers internally, you would rent 50 servers in the cloud for the full 30 days a month. In reality, the rent by the hour approach means you could be renting an equivalent of 1/2 to 1/10th the amount.
2. Elasticity for Unmet Peak Load: How does a CIO provide a 50 node cluster for one small group in the company that wants to analyze last 17 years of demographic data? By saying "No, we can't do that. It will cost $7 million." With an external platform, he can say yes, and pay for 50 nodes for 3 hours.

These two apply for Cloud infrastructure providers like Amazon Web Services. There are additional economies of scale that kick in when you look at multi-tenant application platforms such as With a multi-tenant platform, you are now talking about having a highly optimized environment where multiple customers are leveraging the same set of servers (and cost of maintenance, upgrades, security, etc.). The McKinsey duo does make a distinction between Cloud infrastructure and Cloud services but omits any mention of the advantages of moving the applications to Cloud Service platforms (in their jargon).

Credit Where Credit is Due
As a blogger, it is my duty to throw gasoline on a burning fire and not rain on the cloud parade! Pardon my puns. But to their credit, the McKinsey duo (which I assume does not reflect opinions of all of McKinsey) does make some points that I agree with.

First, the private cloud is not a magical solution to your data center challenges. They suggest - Rather than create unrealizable expectations for “internal clouds,” CIOs should focus now on the immediate benefits of virtualizing server storage, network operations, and other critical building blocks.

Second, virtualization is a powerful tool for making your existing applications in the data center use fewer resources.

Third, public clouds are providing business benefits to SMEs. I concur. But I believe and large organizations like Cisco, Dell, GE, etc. that have been cited for using SaaS applications are proof in my view negate the notion that cloud computing is somehow not enterprise ready. A look at the customer list of vendors like Google,, Taleo, Workday etc. can prove otherwise.

My Enterprise Irregular friends and other bloggers have already pitched in with their opinions.

Deal Architect's Vinnie writes - I would normally ignore yet another “overview” of clouds, but being McKinsey it will get read by executives and several of their generalizations about “not being cost effective for large enterprises” are just plain misleading.

Appirio's Balakrishnan maps out the Cloud with infrastructure, platform and applications as three different tiers with their own pros and cons. He also questions McKinsey report's primary topic - moving existing apps to the cloud, and concludes with - We have seen the benefits of cloud platforms first-hand at over 150 customers, including companies like Avago, Genentech, Japan Post, Qualcomm, Starbucks and Dolby. Once customers experience the benefits of cloud platforms - quantifiable savings, rapid time to value and innovation that drives the business, they seldom want to go back. This is why 90%+ of customers plan to increase their spending on cloud platforms. In these economic times, there is no greater vote of confidence for cloud platforms than that!

Nick Carr is much more postiive on the report but even he finds that the report entirely ignored SaaS applications - 'The cloud also, of course, provides a way to tap into powerful software-as-a-service applications that can provide substantial savings, not only in equipment and labor but in licensing and maintenance fees, over the cost of installing an in-house application. (The McKinsey analysis ignores those opportunities.)'

I expect the blogosphere to continue this discussion and I hope the discussion will include the missing pieces from the report - benefits of Cloud Platforms and Cloud Applications - and not just focus on infrastructure Clouds.

Google has an excellent post on this titled 'Official Google Enterprise Blog: What we talk about when we talk about cloud computing'.

2 comments: said...

When you pitch for's platforms or SaaS, could you please throw in your disclaimer that you earn your living through those?

Anshu Sharma said...

I think its well known to my blog readers that I work for (and that I opine as an individual here - per my disclaimer.)

Your comment would be more meaningful if you did not anonymize yourself by using other company's brand names as your name.

But, yes, your broader point is well taken.