Attempts to address questions of private vs public cloud adoption necessarily involve detailed examinations of the economics. Obscured in these discussions, however, is the influence time has on perceptions of the economic costs and the theoretical return. Current return versus future costs is an important equation, and how it is parsed depends heavily on the availability of local resources.
The fundamental value proposition of the public cloud is accessible economics, as we discussed with Marten Mickos recently. As Infochimps’ Flip Kromer puts it, “EC2 means anyone with a $10 bill can rent a 10-machine cluster with 1TB of distributed storage for 8 hours.” The pricing established by the original market entrant, Amazon, have served two purposes. First, they have made resources available to even individual developers that would have been out of reach absent the pay per use model. Second, and perhaps more importantly, Amazon imposed a ceiling on cloud pricing; a ceiling which they continue to lower as they’re able to leverage larger economies of scale. Whatever the larger ambitions of other systems vendors with respect to the cloud market, Amazon defined the context in which they all must compete. At a pricepoint sufficiently low that the majority have chosen not to.
Developers have flocked to Amazon’s platform [coverage] not because it is the lowest cost option, but rather because it is the most accessible. From an economic perspective, they’re trading up front capital expenses for potentially higher ongoing operational costs. This trade is most attractive when resources – hardware and people – are few. When a developer can’t outsource the task of hardware acquisition and setup, and wouldn’t have the hosting facilities even if they could, the cloud is a compelling alternative.
It is less apparent, however, to larger entities that the public cloud economics compare favorably to their internal operational costs. Setting aside the question of whether their conclusions are sound, discussions with those implementing private clouds as an alternative to public implementations focus on the sustainable aspect of the economics. The accessibility of public cloud services is of less value to larger institutions, both because their expectations in terms of deployment speed are very modest and because they are typically not resource poor with either available hardware and IT resources.
Instead, larger enterprises focus on the operational margins of public cloud. Amazon’s pricepoint has historically commanded a margin above the cost of traditional hosting suppliers [coverage]. Given that larger enterprises generally argue that they can deliver infrastructure at a lower cost than traditional suppliers, the economics tilt even more strongly against large scale public cloud implementations.
The benefits of the public cloud remain of interest, however. What enterprise would not want a more elastic, easily provisioned infrastructure? Private cloud is the inevitable compromise. Promising feature benefits similar to those available on public infrastructure but with what is perceived by enterprises to be a more sustainable economic model, the private cloud is an increasingly attractive proposition for large enterprises.
Supporters of one or the other approaches may question the substance of the above characterizations, but these are the behaviors we have observed repeatedly. If you’re selling private cloud solutions, then, you’d do well to understand the appeal of accessible economics. Conversely, public cloud vendors may want to more clearly articulate the sustainability of their economics over time.
Disclosure: Amazon is not a RedMonk client.