September 4, 2012
In the last little while I’ve seen a couple of posts that focus on the difficulties around calculating total cost of ownership of cloud versus on-premise infrastructure. First up on TechCrunch was a guest post from Jinesh Varia from AWS who correctly pointed out some of the difficulties around calculating TCO and admonished organizations to ensure they included non-direct costs when calculating TCO. In particular he pointed out the personnel costs of running on-premise which are often not included when making TCO calculations. He also correctly pointed out that usage patterns have a direct bearing on TCO. As he says;
If you do a TCO calculation assuming a set peak capacity for your workload, when in fact it has a spiky usage pattern, such as a flood of web traffic during a sporting event, then you might not get the right cost picture in your TCO analysis. When determining TCO, you should consider the nature of the application and historical statistical data to help determine the usage pattern.
Following on from Varia’s post, and indicating this is an area that Amazon are very keen to focus attention on, came a post from Amazon CTO Wener Vogels about TCO and what he calls “return on agility”. The TCO part of Vogels post largely reiterates what Varia said in his post, and alos pints to an AWS whitepaper which includes some interesting metrics around total costs for workloads showing different usage profiles. In the whitepaper AWS makes some comparisons with estimated on-premise costs (both direct and indirect). Obviously the comparisons are a little self-serving but in general they point to the lower cost that cloud delivers when compared on an “apples with apples” basis with on-premise.
The second part of Voegls’ post deals with an area that is, in my mind at least, more interesting, return on agility. In his post Vogels states that;
Many of our customers come to AWS with a reduction of TCO and other cost savings in mind but after using AWS for a while most of them will claim that agility is the even more benefit for them. Agility has many different faces within an enterprise : Operational advantages such as setting up infrastructure in minutes rather than months, completing massive computational projects with a large number of resources quickly, and scaling architecture up and down to provide the needed IT resources only when you need them, deliver targeted IT solutions fast for individual business units – these deliver a “return on agility.” The return on agility delivers business value by allowing you adapt rapidly, while remaining focused on your core competencies rather than distracted by operating IT infrastructure.
But more importantly it allows the business to change and dramatically reduce time-to-market for products. It drives down the cost of experimentation and allows companies to explore many more product directions with minimal risk. In the current economic climate where capital is scarce, being able to develop new products rapidly without the need for major capital investments is crucial to the success of businesses.
It’s a topic that I’ve been waxing poetic about for a long time now. While it is undeniable that in the majority of cases cloud will be cheaper than traditional models of delivery. The benefits that cloud brings in terms of agility and flexibility far outweigh the cost benefits – looking at TCO alone is a race to the bottom of the cost-cutting hill.
What both of these posts fail to cover however is the dearth of options organizations have to really measure what is important for them – the economic impacts of the agility that cloud is delivering for them. What I’m arguing for here is a solid econometric model that goes beyond TCO and delivers an accurate return on investment (ROI) measure for the cloud – more on that in part two.