McKinsey's Cloud Computing Report Is Not So Clear

By Leena Rao

McKinsey & Company released a report, “Clearing the Air on Cloud Computing,” Wednesday that claims that large corporations could lose money through the adoption of cloud computing. 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.

The report makes some thought-provoking points but neglects to address a few key trends that are occurring in cloud server services. Innovation is rapidly changing in the cloud. The space is still very much a work in progress and big cloud computing services, like AWS, Google (GOOG), Sun Microsystems (JAVA) and Microsoft (MSFT), are regularly coming out with different products. As these companies throw their hats into the “cloud computing ring,” AWS will face increased competition in the market and could cause prices to go down to fight for market share.

Amazon’s (AMZN) cloud computing services, in particular, are constantly evolving. What started out as pay-by-the-drink storage (S3) and computational processing ((EC2)), now includes a simple database (SimpleDB), a content delivery network (CloudFront), and computer-to-computer messaging (SQS). Most recently, Amazon added a web-scale data processing engine with Amazon Elastic MapReduce. (It is a framework for accessing data stored in file systems and databases). It allows developers leverage Amazon’s cloud computing power by creating applications which process huge reservoirs of data (conveniently stored in Amazon S3) in parallel.

The next generation of enterprise apps is already begun to be written with the cloud and virtualization both in mind. At that point, it doesn’t make much sense to do it all through conventional data centers, when you can optimize through other services and get the best of both worlds. And many large companies currently use cloud services for a segment of their data storage, and also utilize and virtualize in conventional data centers.

Microsoft recently announced Exchange 2010, a new suite of Microsoft Office-related products, are designed to be deployed and managed servers on-premises or from the cloud. Microsoft’s Azure OS, which is expected to be rolled out in the fall, can host these office-related products in the cloud. It’s not a huge stretch of the imagination to speculate that in the not to distant future, Microsoft will integrates the on-premise storage and Azure storage together, thus allowing companies to tap into both utilities in the same application.

The report seems to hype the cloud costs and understates the rapid changes in cloud market conditions and resultant innovation and price cutting that will take place in the near future.

Original post

Comments

[raytayzmd]

...as I posted elsewhere: who paid for the report?...it looks like it was for the "Uptime Institute"...from their website:

"Uptime Institute Professional Services, formerly known as ComputerSite Engineering, Inc, is the data center engineering and management consulting group of the Uptime Institute. Our mission is to address technical aspects of contemporary data center issues."

...hmmmm -- so I guess "cloud computing" success could be bad for "Uptime Institute" business...and, gee, McKinsey's report warns us of the evils that may befall "overhyped" victims...well -- surprise, surprise!...but never fear for there is a solution:

"Users, hardware vendors and service suppliers can take
specific steps to ensure the successful adoption of cloud
technology—and prevent it from getting stuck in the “trough of
disillusionment”

...and, by golly, five'll getcha ten that with a little help from "Uptime Institute," you and yours 'll be able to sidestep any misfortune that might develop from messing with clouds...sounds like just another advertisement cloaked behind a veil of "authority" -- "authority" for a price, of course.

2009 Apr 16 05:33 PM

[David H. Deans]

@raytayzmd I agree that this assessment reads like the kind of "analysis advertorial" that Aberdeen Group will produce for the right price.

Frankly, even though I'm sure that McKinsey & Company has suffered a loss of consulting work due to the economy, I'm puzzled why they would risk damaging their brand a associated reputation. Very short-sighted.

2009 Apr 17 11:06 AM

[cloudberryman]

Cloud Computing is changing the way people do IT in any size organizations. One of the things that should help in adoption of Cloud Computing for businesses is the right tools such as our CloudBerry Explorer for Amazon S3 cloudberrylab.com.
I also believe that large companies, especially on the department level will think twice before buying new hardware/ software and may favor cloud-based services instead.

2009 Apr 19 06:14 AM

[Unni Narayanan]

McKinsey has utilized a “cloudy” methodology to oversimplify data center design challenges and “rain” on SAS’ parade. In reality, CIOs of large enterprises are focused on a combination of solutions to optimize metrics such as TCO (Total Cost of Ownership), so there is no “one size fits all” solution. This means that cloud computing cannot be summarily dismissed, nor is it a panacea to persistently overcrowded, overheated, and expensive-to-maintain server rooms. It’s more complex than that. First, inherent problems in the data center are accelerating a push toward co-location and managed-services providers. This provides a fertile backdrop for cloud computing to flourish:

(1) Power consumption and the associated cooling problem play a significant factor in TCO. Although, a silicon valley CTO may enjoy 95 degree heat OUTSIDE of his data center, he cannot allow his data center to burn that hot. If he does, he will find himself toasted to a crisp and out of a job. So, CTOs are already riding the co-lo wave because they cannot afford the capex to pay for decent air conditioning.

(2) Under some models, personnel costs drive TCO even more than the utility bill. The constant pressure to downsize means that application-specific expertise has to move offsite. We at Primary Global Research have observed time and again that personnel cost savings are very real (contrary to the McKinsey study).

(3) Data set sizes are increases and the associated storage maintenance (i.e. applications) are fueling a storage consolidation wave and driving greenfield SAN/NAS implementations in data centers.

Add into the mix, McKinsey’s faulty assumptions:

(1) The benefit of cloud computing comes with the economy of scale. Amazon’s current web services pricing cannot be the basis of a fair comparison.

(2) McKinsey has an oversimplified view that that leasing of HW does not make economic sense. At first glance, McKinsey appears to be correct. In fact, some co-los have leasing models where they charge as much as $100/month per GB of DRAM on a server. However, a deeper look into the matter reveals that (a) HW leasing pricing is still not efficiently priced in the market and more competition will eventually drive costs down (b) Costs are not exactly what they seem. For example, in the case of the DRAM, there is a measureable effect in terms of server power costs related to the amount of DRAM on the system. And so, what looks like an exorbitant server leasing cost may actually be a broader capex issue (in disguise) that the CTO no longer needs to worry about.

(3) McKinsey makes an obvious statement when it suggests that enterprise should aggressively deploy server virtualization solutions. However, the ultimate TCO benefit from virtualization may arise in the cloud computing environment. In fact, we at PGR have not yet seen a vendor that has provided a viable “virtualized” stack in a managed-services environment; this is probably a ripe business opportunity.

(4) Although, McKinsey was focused on large enterprises, the real volume of customers may arise from the SMB end market.

I am not saying that all Fortune 100 data centers will be “living on Cloud 9.” Certainly that will not be the case. I am saying, however, that over the last 50 years there has always been a natural tension between the move toward centralized versus distributed computing resources. Haven’t we all repeatedly heard the debate about whether the mainframe was going to die? Unlike McKinsey, I think it is too early to predict the death of cloud computing. I definitely see the silver lining.

Unni Narayanan, Ph.D.
CEO and President
Primary Global Research, LLC

2009 Apr 23 04:43 PM