Recently, McKinsey released a report titled “Clearing the air on cloud computing” and it has upset many of
the Cloud Computing advocates. According to this report, the cloud computing
services like Amazon EC2 might offer substantial savings for small and medium
companies but they are pretty expensive from the enterprise point of view. They
also argue that enterprises can cut costs by using suitable virtualization
technologies. As expected, this report is being used by those who are threatened
by Cloud Computing to whip up FUD.
I cannot dismiss the report entirely but I have some problems with it. If you
check the slides on page 22 and 23, they have made some cost comparison between
Amazon EC2 and the traditional enterprise datacenters. This is the part I am
uncomfortable with in the report.
- They have quoted the Total Costs of Assets (TCA) in the case of traditional
datacenters to be $45 per month. This seems to be extremely low amount from the
back of the envelope kind of calculations I made. - I am really not sure if they took into account the total energy consumption,
cooling costs, etc., in their calculation. Unless I am doing something terribly
wrong, I don’t see how they can get such a low TCA. - To me, it appears that they have not included the labor costs associated
with hiring technicians and others to manage all aspects of the datacenter. - Also, one of the advantages of Cloud based servers is their utility pricing.
I am not sure if they took this into account while calculating the costs
associated with the cloud server resources and usage to match the costs
calculated at the 10% utilization of the traditional datacenters. The resources
on the Cloud need not be considered to be “on” 24/7/365 like the traditional
datacenter resources while doing the math for the costs.
I have some serious misgivings with the numbers presented in the report. Here
is my request to all the analysts who throw up numbers to justify their
conclusions. Please present the damn calculation in its entirety. In the absence
of such an information about how the numbers were crunched, there is no way one
can prove or disprove the conclusions. This leads to further confusion and the
associated FUD.
It is not just me who is feeling about the report this way. The folks at Rightscale also think that there is something messy in their calculations
In particular, it doesn’t seem to be accounting for the costs correctly and
it completely fails overlook the benefits of automation in the cloud which
ultimately leads to a revolution in the way compute resources are consumed.The cost equation in the
report starts on slide 22 and it’s really, really sketchy. They mix EC2
compute units and cores together (compare 22 and 23). They talk about
“$14K/Server (2 CPU, 4 core)” which on my calculator comes out to $97/core/month
over 3 years, but they have a cost of $45/mo/cpu on the same slide (and $97
doesn’t even account for the facility or power or cooling).On slide 24 they suddenly compare an in-house datacenter server with “75% of
EC2 Large Standard Windows configuration on Amazon EC2″ and nowhere do they
mention that the latter cost includes the Windows license.
Ouch!
I just hope that the author of the report follow it up with a blog post
explaining how they made their calculations. If not, this report means nothing
except for its FUD value.
I will also take this post to vent my gripe against Amazon’s pricing
structure. In my opinion, the pricing is still too high for something that is
supposed to be commoditized like electricity. I was hoping for the prices to
come down but so far there is no movement on that front. There is some
competition coming up from GoGrid and Rackspace but, at this moment, their
pricing strategy is not a cause of concern for Amazon. If Amazon is serious
about getting into the enterprise IT with full force, they have to lower their
prices drastically. Unless the pricing is very attractive, enterprises are not
going to get out of their comfort zone anytime soon.
Hm… what is it with questionable McKinsey studies?
Quite some time ago I debated their study which predicted Accounting to be amongst the last to switch to SaaS.
They might want to read Ben’s Accounting 2.0 series 🙂
Well, of course AWS has the same cost inputs as an enterprise Data Center – power, support, security , hardware, etc. There is no ‘magic dust’ that AWS or any other cloud provider has that allows them to produce computing at a lower unit cost than a large enterprise could or even traditional IT services supplier, using the same technology.
Clearly AWS is leveraging scale, high levels of virtualisation, and automation to push the unit costs as low as they can go. But so can a large enterprise IT shop – and they would not have to pay for AWS profit margin and overheads of cloud management.