Grounded in a survey of 53 Infrastructure as
a Service (IaaS) providers, 451 Research has released its 2013 Cloud Pricing Codex, which is designed to give users of cloud services (enterprise CIOs and CFOs included) a road map of how to more fully understand IaaS offerings.
The survey found that pricing models vary greatly among IaaS
companies; only 64 percent of providers even publish their pricing online. But
authors of the report, Owen Rogers and William Fellows, assert that all virtual machine pricing methods fall into eight
classifications. They call this a cloud pricing taxonomy, and they’re
confident that the Cloud Pricing Codex can simplify a complex cloud economic landscape for IT
TechRepublic recently had a conversation with Cloud Pricing Codex
co-author Owen Rogers. Our talk focused on the key findings given in the executive overview of the report.
Before starting at 451 Research this past year, Rogers earned a PhD in cloud economics at the University of Bristol. That got my attention—I asked Owen if he knew of anyone else with such a PhD
specialty, and he said he didn’t know. And so all things cloud continue to
Key takeaways from the interview:
- Main question: What sorts of things are we
buying with cloud services, and how much do they cost?
- IaaS providers use different units of
measurement and different pricing schemes. It’s hard for consumers to know exactly
what they are getting.
- Rogers found it impossible to compare individual
cloud providers to each other. Over time, he saw the commonalities between them,
and this lead to the classification taxonomy in the Cloud Pricing Codex.
- The aim of the Cloud Pricing Codex is to remove confusion so users can
make reasoned judgments about cloud services.
- The taxonomy provides a way to understand the benefits and the financial risks of IaaS provider plans.
- Rogers hopes the Cloud Pricing Codex will give users the “most
acceptable understanding” of the options and risks of IaaS.
- Cloud users begin experimenting with services and then shift to alternative pricing models in order to lock in necessary
resources and reduce cost.
- Without knowing the details and the risks, a virtual
machine can become a component of greater costs as more cloud resources are
- A digital infrastructure strategy is really how
a business sees the cloud working for them. The Cloud Pricing Codex is the first check in
understanding the options.
- 451 Research plans to continue surveys for the
Cloud Pricing Codex; they expect an update by the end of 2014.
TechRepublic: Let me ask a question posed in your report.
Why is the Cloud Pricing Codex needed?
Owen Rogers: I think the simple question should be: what
sort of things are we buying with cloud services, and how much do they cost? It
is a simple question to come to, but in reality it is quite a difficult question
There are a number of reasons behind this. First, different
cloud providers use different units of measurement. Different cloud providers
offer different services and have a range of different pricing schemes as well.
Cloud providers at first provide on-demand pricing, but there are other models.
And because of all these different
factors it’s really difficult for consumers of cloud resources to know exactly
what they’re getting and know exactly what their cloud bill should be.
TechRepublic: In your methodology
for the survey, you uncovered a lot of findings, which you organized quite well
in my view. You surveyed 53 different IaaS providers. What approach did you use
to uncover this level of detail? I think it’s going to be very useful to anyone
who reads the report.
Owen Rogers: That’s an interesting question, actually. First
of all, when I started the survey I contacted the service providers and asked
them for all the information they could provide. The amount of detail that I
got from the providers was really very generous.
It quickly became clear that understanding the different
pricing methods among the providers was going to be an almost impossible
challenge. So rather than attempting to compare individual providers to each
other I started to think about where providers had commonalities. I used that
as the basis for identifying the eight standard pricing categories in the IaaS
When I started this survey there
was no way to compare which pricing method Provider A had compared to Provider
B, and this is really where the whole concept of the Codex came about.
When you first look at the IaaS market, it seems to be this
confusing array of pricing models and different ways of doing things. But
actually when you collect all the information and absorb it over a long period
of time, you begin to see that cloud providers use similar methods to represent
what they’re doing.
The aim of the Codex really was to remove the veil of
confusion so you can make more reasoned judgments on what you’re likely to pay.
The Cloud Pricing
Codex gives a taxonomy (see Figure 1)
of eight pricing classifications, under
four separate categories. They are: Cash Pay-As-You-Go, 1) On-Demand 2)
Spot Pricing 3) Reserved Instance; Committed VM, 4) Prepaid VM 5)
Recurring VM; Resource Pooling, 6) Recurring Resource Pooling; Credit
Pay-As-You-Go, 7) Subscription Credit 8) Prepaid Credit.
TechRepublic: Let’s say I am a decision maker at a
reasonably-sized enterprise. If your Cloud Pricing Codex taxonomy is like a road map,
how do I take this map into my hands and use it effectively?
Owen Rogers: I think there are two primary ways of using the
Codex. The Codex lists the results of the survey—what cloud providers and which pricing methods they use. Anybody who buys the report can see exactly what pricing
methods a provider uses, so that they can remove all this confusing terminology
and look at the bare bones of the provider’s offer.
I think the second reason is that the taxonomy provides a
way of understanding exactly what the financial risks and benefits are. The
Codex goes into greater detail about each of the pricing methods and talks about many of the risks. So for example, you may
have an element of risk if you have to pay up front to secure a small price per
hour opposed to on demand, which is completely risk-less.
I’m not sure the market is
completely aware that all these various pricing methods have their different
benefits and different strengths. I think there are some applications where
these different pricing methods may look better and other applications where
I hope the Codex gives consumers the confidence to really
understand what options are available to them and use it to assess what the
best value and the least risk is for them. Perhaps a better way to say it is
the most acceptable understanding of what the risks are.
TechRepublic: Looking at Figure 1, I wanted to understand this category better—resource pooling. Could you give
a quick case example of what resource pooling is in infrastructure as a
Owen Rogers: That is similar to the committed bandwidth
check model. So a consumer every month or so would commit to a certain amount
of CPU memory storage. Then through the month while they are consuming that,
they benefit from that prepaid amount. But if they need to use more than they
budgeted for, they can buy additional resources in real-time, but that is
charged at a higher price than the resources you paid for in advance.
TechRepublic: So you’re doing well as long as you stay
within the limits of your plan—much like a consumer plan with a cell phone. If
you go beyond that, you’re going to incur higher risks and costs.
Owen Rogers: Higher costs, yes. If you follow the plan,
you’re paying the lowest price. But the benefit of the cloud is, if you end up
using more resources than you budgeted for, you can scale up and still consume
those resources. But they will cost you more.
TechRepublic: So you’re kind of trading scalability for
Owen Rogers: Yes. It’s this whole kind of concept—are you
very predictable, or are you very unpredictable? And if you are predictable, I
think you are more likely to secure these kinds of discounts.
TechRepublic: The second key finding of the Cloud Pricing Codex report
states: “On-demand pricing enables experimentation; alternative pricing enables
cost-effective implementation.” Consumers start out experimenting, but then they
want to get to more predictability, which IaaS companies provide through
alternative pricing. What was the process that you found?
Owen Rogers: I think the general situation is that cloud
computing, because the costs are so minuscule, it’s
easy to get started—you can pay with a credit card. That encourages users to
just sign up and start playing around with different cloud providers. I know
myself, I signed up for different cloud providers and the cost is just a few
cents every month. With that I get to experiment and play around and see what’s
interesting in the market.
Over time I suppose as use cases
of the cloud mature, consumers are likely to spend more on-demand and add more
virtual machines, and then slowly but surely see their bills slowly creeping
up. And I think it’s a bit frightening when they start using the cloud for
their enterprise applications, for things that are mission-critical.
They start thinking to themselves,
well I wonder if there’s a way to continue to have the benefits of cloud
computing with the scalability, but that we can also get discounts to have
access to more resources in the future. And this is where the alternative
pricing methods come into play.
If you have some idea about how
many cloud computing resources you are going to use in the future, you could
lock in that price so that you already have access to the minimum requirement.
But you can still use the benefits of on-demand to purchase additional
resources for the unexpected.
TechRepublic: Your third
key finding from the report overview is an interesting statement: “A virtual
machine does not an application make.” I will let you expand on that very
attention-getting statement. It seems this refers to a reality, if I’m correct,
where you don’t have a one-to-one correspondence between a virtual machine and
an application in the traditional IT sense.
Rogers: Yes, I think that description is spot-on. Having a virtual machine with
an application stack on it—I suppose you could call that advancement of cloud
computing, but I would not necessarily call that a cloud application.
For me a cloud application utilizes the benefits of what
cloud affords. Any application that is built using the cloud should be able to
scale on-demand. It should be loosely coupled, and should be able to take
advantage of different data sources.
In terms of cloud pricing, of course, you have to take
account of the cost of the virtual machine, but you also have to take account
of the cost of other things as well, such as bandwidth, external storage, and
backup transactions. And there are all these other small cost items that I
think many people forget about.
so focused on the virtual machine, they think all these incidental costs won’t
really matter. But obviously the more storage you use externally, and
the more customers visit your website, there might be in an increase in
bandwidth. And slowly and surely these little charges will creep up, and the
virtual machine will be a component of a much bigger unit of cost to worry
TechRepublic: Your fourth finding in your overview refers to
creating a digital infrastructure strategy, or a “playbook” as you call it. To
use a sports analogy, how would an end user with the benefit of the Cloud Pricing Codex take
the playbook and create a winning record—create more benefit and value in using
Owen Rogers: So I think the Codex is really a third step in
a complete understanding of the cost implications of the cloud. We believe in
this concept of multiple cloud venues—the idea that in the future consumers
will need to use multiple clouds for different purposes, and they will choose
the best cloud at the time based on their physical needs and cost and location
So the strategy is really the way the business sees the
cloud working for them. It’s how the world of cloud will contribute to the
business, and how IT will be a fundamental part of scaling the business and
entering new markets.
The playbook and the digital infrastructure strategy—these
are the ways of understanding how the clouds and the business link. The Codex
is kind of the first check: what are the risks of doing it this way? Should I
do it internally or externally? What types of pricing risks are available to
TechRepublic: It occurred to me while I was reviewing your
materials that cloud computing is still very new. People are still catching up
to it in the business world. I am wondering whether 451 Research is planning to
update the Cloud Pricing Codex and come up with further editions as trends and conditions
Owen Rogers: Yes, we will do that. I think you’re right—the
market is in a really interesting stage because it has been constantly evolving
for quite a while. I wonder now if we are going to see new pricing models come
out, or if we are going to see some kind of standardization of pricing models.
I think it’s crucially important that we update the Codex to take account of
these changes in cloud pricing.
TechRepublic: Does 451 Research
have any solid plans or expectations at this point as to a further edition of
the Cloud Pricing Codex?
Owen Rogers: We don’t have any specific plans at the moment, but I think we can be fairly confident that in the second half of this year
there will be some kind of update.
TechRepublic readers can review Owen Rogers’ analyst reports on the 451 Research website.