Budget issues are biggest driver for cloud services: IDC

The benefits commonly ascribed to the 'cloud' or on-demand model, in order or importance, are "easy/fast to deploy", "pay only for what you use", and "offers the latest functionality", according to a recent IDC research.

These three benefits out-rank others, including "less in-house IT staff, costs" and "low monthly payments", said Claus Mortensen, who cited the findings from the IDC Asia Pacific End-user Cloud Computing Survey released in January 2009.

Mortensen is the principal for IDC's emerging technology advisory services, including the digital marketplace and new media program, based in IDC's Asia/Pacific headquarters in Hong Kong.

Defining cloud services

IDC defines cloud services as "consumer and business products, services and solutions delivered and consumed in real-time over the Internet". For example, cloud applications, cloud collaboration, and cloud IT management can be grouped under the software-as-a-service package; cloud platforms can be delivered as platform-as-a-service; while cloud storage and cloud servers/processing can also be delivered as infrastructure-as-a-service.

The hosted models of cloud computing can be broadly separated as 'public cloud computing' and 'private cloud computing'. Public cloud computing are best suited for data centres that are dedicated either for production or development purpose. Examples of applications are custom applications on Amazon Elastic Compute Cloud dedicated websites, IBM applications on demand, Fujitsu trusted-service platform, SalesForce.com, Oracle-Siebel CRM on demand, MS Live CRM, Force.com, and Google Apps. As for data centres that have dedicated applications such as financial systems, databases of healthcare and insurance, private cloud computing services are recommended, said Mortensen.

The same research indicated that cost and budget issues are the main reasons that organisations are driven towards using or considering cloud services. About 52 per cent said the need to cut cost was the main driver, another 28 per cent said they were facing budget issues which force them to find cheaper alternatives for data centre management.

Data centres carry three main performance metrics that suit different types of needs. The present recession has changed the organisations' decision on where to hold their assets, said Mortensen. Type A is enterprise data centres, which are most concerned about meeting service-level agreements and the efficient use of space. Type B is hosters or outsourced data centres, which place the greatest emphasis on price, performance and scalability. Type C are data centres used by Web 2.0 service providers or telecommunications company, which pay the most attention to density, energy efficiency and the time to market. While Type A data centres tend to require fixed pricing schemes, they require the most capital expenditure, followed by Type B and then Type C. On the other hand, while Type C data centres tend to offer variable pricing, their costs are off the books as they only require an investment in operational expenditure.

Blueprint for future data centres

Cloud services end-users have also indicated their perceived level of use of cloud services in their organisations in the same IDC research. The future use of cloud services, in order of importance, are network capacity, storage capacity, management applications, server capacity, collaboration applications, followed by business applications, application development and deployment, and personal applications. Compared to the current level of use of cloud services, the use in collaboration applications has recorded the largest increase, while the use in infrastructure services (network, storage and server) has shown up consistently well.

The blueprint of future data centres, said Mortensen, can be summarised into three key attributes. First, future data centres will adopt a modular and design which can be repeated, which will enable IT to reuse data centres locally, regionally and worldwide. Second, data centres in the future should be planned with scalable capacity. This will help reduce over-provisioning and will serve their purpose just in time. Finally, data centres should be operating in a predictable manner, as familiarity drives higher levels of availability.