"The Cloud Conversation" for Every CFO

If you haven't sat with your CTO or CIO over "The Cloud Conversation", then now is the best time to start considering this especially when your enterprise is already beginning in its transition.

The cloud that needs to be discussed is the infrastructure cloud - the kind of cloud Amazon has pioneered with its Amazon Web Services (AWS) offering. Amazon and friends are changing the economics of IT infrastructure through the application of scale and specialized expertise in much the same way that Amazon.com changed the rules for the smaller book store.

There is a whole separate conversation to be had concerning the agility, flexibility and transparency that comes with a cloud infrastructure - but we are going to try and focus here on the financial and economic advantages and prepare you for "the conversation".

The big picture is that there are three economic legs to the cloud infrastructure stool: Utilization, Specialization and Scale.

Scale -- The scale advantage is simply that Amazon, Microsoft, Google and others are investing billions of dollars in building out state-of-the-art data centers and then renting usage back to customers by machine hour - the price point has already come down as low as a few pennies per hour - so you get a small piece of their buying power and because you only pay for actual usage, by the hour, no CapEx is required.

So the first response from your CIO might be that your current hardware costs equate to similar or even fewer pennies per hour when costs are amortized over 3 years, even after cost of capital is included - and they are probably right - the cloud providers are adding a margin back into the price they charge you!

Specialization -- BUT, and here's the second leg of the stool - servers and storage are probably only about 50% of your IT delivery costs (and I'm excluding the cost of the DC build out and the cost of people). Power (and the equipment required to keep your hardware cool) is the other major cost component and it has been growing over recent years.

In older, less efficient facilities, this can be a larger cost that the actual servers and storage. If you are using a collocation facility, you are probably very familiar with the costs I'm talking about here.

The infrastructure cloud providers are at the forefront of innovation to increase energy efficiency and reduce this cost - they are quite probably between 50% and 100% more efficient that your current environment today, and investing substantial sums to get better.

The bottom line is that after combining hardware and power/cooling costs, you are likely already paying more than the price available from an infrastructure cloud. And we still did not even mention about the very important advantage of the infrastructure for the cloud.

Utilization -- The third leg of the argument, and the one that uncovers significant advantages for many, is a result of the hourly usage pricing model. Most data centers operate at 10%-20% utilization, some as low as 5%. This means that at least 80%-90% of your infrastructure is sitting idle for most of the time - adding no value, just costing you money. This extra capacity is there for those few weeks, days or even hours when you hit peak demand - end of year sales, end of month processing, major promotions etc.

Simply by moving to a usage-based payment model in a cloud that scales resources up and down as required, you will eliminate most of the cost of idle capacity. This would have implications for your company which would largely depend on your present use and the variables of the demand, though it's not that uncommon to see at least 50% of savings on costs as soon as the applications are moved from a traditional data center or hosting provider to an infrastructure for the cloud.

At this point in the conversation, if not before, you might hear from your technical team that the public cloud infrastructure is too risky and that there are security concerns.

But the main points I would suggest are that (a) the object here is to start down a path with a pilot project, not start with the crown jewels, (b) there is already a ton of money and many companies focused on addressing these issues to a level well beyond their resolution in most internal data centers today, and (c) there are many very large and conservative organizations who have found enough here to start down their journey.

Another response you should be prepared for is that the team is already well ahead of you and evaluating a private cloud strategy. This is great news.

But private clouds have a different economic model - they involve you buying the infrastructure and delivering on demand computing resources across your business. So the business advantages of responsiveness, agility, and transparency should be the same as for a public cloud like AWS, and your team will have greater control of the underlying physical infrastructure. But you may not get the economic advantages of scale, specialization and utilization we have been talking about here.

The traditional infrastructure vendors are very keen on this private cloud model - for the obvious reason that they can still sell you all the same kit. This could be a great solution for your organization, especially if you are very large and can consolidate IT infrastructure resources across divisions or businesses. It will not automatically deliver the cost advantages spelled out above. My recommendation would be to push for a parallel project to try out the public cloud and provide context and comparison for private projects.

We hope this helps equip you for the conversation. Cloud infrastructure is not a silver bullet. Like every other new thing, it has its own set of challenges - though it is very necessary and will dramatically change your costs for IT and that of the competition.

Now is the time to ensure that your IT team is engaged and gathering experience and perspective to build out a plan that is right for your organization.