Get your payment model right

It is easy to see it is early days yet in the implementation of cloud services, despite cloud moving closer to centre stage as a contributor to the delivery of IT resources to business. This is particularly clear as users start to count the actual costs of using cloud over their first year or so of operations. According to hosted services provider, Claranet, the wide range of payment options available to users is causing not only confusion, but also examples of over-payment.

For example, the company’s third annual cloud adoption survey has thrown up anomalies between pay-per-use and fixed rate payment models that may catch out the unwary.

The survey polled 300 IT decision-makers from a range of small and medium-sized businesses and enterprises. It found that 34 percent of large organisations (those with more than 3,000 employees) that are using cloud-based services are taking these services from more than one provider and are tied to more than one payment model as a result. Payment models cited included flat-rate subscriptions and a variety of scalable and pay-per-use options.

According to Michel Robert, Managing Director of Claranet UK, the apparent flexibility offered by pay-per-use models may give the impression of value for money, but the perception should be treated with caution.

“Flexibility has always been one of the key selling points of cloud computing and outsourced IT, and this extends to the models available when it comes to paying for services. Just as they need to ensure the service they take is suited to their needs, organisations must make sure the payment model they choose is appropriate to the way in which they will use their cloud services,” he said.

“Our research revealed that 87 percent of organisations surveyed cited flexibility of compute resources (ability to scale up or down) and access to applications as a key objective when migrating to the cloud. Yet 75 percent said that their compute usage was predictable. A Pay-As-You-Go model may provide value for money if you are frequently using the ‘burst’ facility of your cloud service. However, if a company’s particular workload is largely predictable it may end up paying more than it needs to. Yet in most circumstances, it should be possible to anticipate your requirements and put in place a payment model that reflects this level of productivity and delivers genuine value.” 

The underlying issue here is that the potential of the cloud to offer almost infinite scalability is an extremely attractive prospect for end users, as they can reduce their dependence on rigourous pre-planning of workloads and the resources needed to service them. But the ability to provide scalability means that service providers need to have the resources available to provision it on-demand.  

And while those resources can be shared amongst a customer base, saving them the expense and time taken to provision new systems, carrying that cost falls on the shoulders of the service providers. And it will get passed on to the customers opting for a pay-per use option.

But as Robert observes, many applications and processes vary little in their resource requirements, which can be predicted well into the future. This does mean the customer has to continue to make that rigourous assessment of resource requirements, and apply the results to a number of different options.

For example, if the application resource requirements can be predicted out for several years, the best option may even be to opt for a bare-metal, dedicated server solution managed by the CSP, rather than a fixed rate payment schedule for a service running on a currently available virtual server.

It would appear from this survey that some of the cost issues around cloud may have been over-hyped, creating a degree of over-expectation in many customers.