There's been a lot of talk about what 2010 represents from a cloud adoption point of view. Some say it's the year of the prototype, others the year cloud computing becomes the standard way people consume web based products and service. For me it represents the year companies innovate on the various cloud business models or to put it another way, it's the year we make money.
As we, [Enomaly] deploy more and more cloud service providers around the globe it's interesting to see the differences among the various regions and classes of services providers and how they adapt our product for their particular market. When we built the latest iteration of the Enomaly ECP Service Provider edition, we decided to make the product as flexible as possible for a variety of different business / revenue and deployment models. This was a risky move, because it now means that we have the extra added responsibility of assisting our customers with not only deploying their cloud services, but also defining any number of various pricing schemes. But as it turns out this decision has given us an interesting opportunity to gain valuable insight in to the various cloud business models.
For example lately a lot of our customers have been choosing to use quota based month subscription models. This model allows a kind of hybrid between a pure utility model crossed with a credit/reservation/overdraft approach. An approach that generates some revenue even if your existing hosting customers never use the cloud capacity. Think of it as a kind of fast cloud recovery or standby. Basically for a month fee cloud customers are given an allotment of potential capacity, say 20 VM's and X amount of storage & RAM. The benefit to this approach is a customer can reserve a certain amount of guaranteed capacity for a small fee and in turn the cost per VM/storage etc is reduced, similar to Amazon's reserved instances. This also fits in well with a lot of hosting companies more traditional existing business models.
The approach provides the cloud service provider with a greater level of insight into potential future capacity requirements and more importantly enhanced revenue predictability. We're also seeing the concept of "cloud overdraft protection" (a concept first suggested by Daryl Plummer at Gartner) where additional burst capacity provided beyond the quota may be charged at a premium. Another potential opportunity of the quota subscription model is it provides a more efficient method for linking multiple regional providers or the so called intercloud. Capacity reservation allows customers to have an efficient and adaptive global cloud recovery strategy with VM's that can sit in a kind of paused or hot standby waiting to be turned on at moments notice, but not consuming any CPU resources. More importantly it provides a sustainable and cost effective reason for regional providers to offer these kinds of services. This approach is perfect for regional scaling and other geographic sensitive requirements. In a sense you're pay extra for the geotargeted luxury.
Another interesting shift we're seeing is that of hypervisor based marketing -- it's dead. Amazon EC2 was among the first to do this but it seems to be catching on more broadly. Rather than marketing based on the particular VM technology say VMware or Xen, there seems to be a shift to "it's just an IaaS cloud" and the hypervisor really doesn't matter. Previously hosting providers seemed to focus on the fact it was powered by ABC hypervisor, but lately we seem to be moving to an approach that is more inline with, "it solves ABC problem" or is based in ABC location. This in itself shows a maturing of the market and more importantly an evolution in how IaaS products and services are marketed.