Last week I was at the VCE Analyst Council in Chicago. What makes VCE different is that they actually build and deliver a complete solution, basically a turnkey datacenter. Everyone else mostly delivers parts at this scale. I’ve been really fascinated about this concept because I think it should be the norm not the exception.
I first saw this concept in action long before I got to know VCE. Back in 2009, Microsoft was pitching a data center in a container. You specified a complete solution (servers, networking, storage) and a shipping container showed up that needed only power, water (for cooling) and to be plugged into the network. Software could be preloaded, and the data center was tested at the vendor manufacturing site, making it amazingly fast to implement.
VCE offers pretty much the same thing without the container and with reassembly inside of your building. In a market that has seen rather dismal growth, VCE has been growing at an impressive rate, showcasing that buyers like this concept.
However, competing with VCE is “The Cloud,” or more accurately, Amazon Web Services (AWS) for the most part. With AWS, time to implementation is also impressively short, and there is no capital cost to deal with. Certainly there are issues with compliance, particularly with regard to security, but often this just takes a change in the policy and some blending of on-premise technology (which likely already exists) into a hybrid model.
The two approaches aren’t mutually exclusive, and current thought is that the ideal solution for most companies should be a blend of the VCE and AWS approaches with dynamic management and compliance assured.
However, there is a tension between the two models, and it isn’t yet clear where we’ll end up.
Let’s assume that both of these approaches can perform the same tasks. The choice then comes down to a couple of things. The most visible is that there is a capital expense to having equipment on-premise and none associated with a cloud service. Many organizations assume that operating expense is better than capital expense, and a lot of firms are structured to make it harder to get approval for capital expenditures than for recurring operating expenses.
But the right way to approach this is by calculating the firm’s cost of capital, because in the end even capital costs can be translated into an expense (in this case interest). The expense of capital is what you pay to borrow or service it (debt or equity financing), and it is always harder than operating expense to calculate and justify. However, interest rates are dirt cheap right now, and Amazon is buying their equipment and marking it up, suggesting that, at scale, buying is likely a lower expense to the firm if it is done correctly. While finance should be consulted about whether the firm buys or subscribes to a service, that is often not the case when a service like AWS is selected, and at least at scale, it should be. (This should be an internal audit review function, and often is).
With the VCE approach, you do get more control over costs, as Amazon can raise rates without asking your permission or giving you any heads up. While the trend for the cost of cloud computing services is down, Amazon is under increasing pressure to improve profits. We’ve certainly seen how, once a dependency is created, vendors have a nasty tendency to raise prices because they know your switching cost will keep you tied to them.
Tactically, AWS is cheap now; strategically, this could become a problem in years to come. Maintaining flexibility so you aren’t trapped would be advisable.