Cloud computing: How to decide 'when to cloud'

Some say cloud computing will become ubiquitous. Others say it's hype and they'd never trust the cloud. Mark Tonsetic, program manager at Corporate Executive Board's Infrastructure Executive Council, says it's not an all-or-nothing proposition. Instead, he advises CIOs to look at the cloud on an application-by-application, project-by-project basis.

Using a weighted scorecard with various factors to consider, CIOs can make a well-reasoned decision about when the cloud is right for a particular situation, Tonsetic told me in a recent interview. For example, the cloud might be ideal for a retailer that needs to supplement internal resources with external capacity when data processing volumes spike during a two-week sales period. Or, the cloud might work for a company that needs fast provisioning of IT capacity in order to reach a time-to-market business goal. The cloud could also work if the internal IT department lacks the skills and enough people to handle a particular task.

Tonsetic shared with me a basic outline for a cloud scorecard:

1. Strategy: Can provisioning via cloud enable competitive advantage (e.g., speed-to-market) for the enterprise?
2. Capacity: Is the workload associated with this project highly variable, such that cloud can provide an outlet for spikes in demand?
3. Security: What is the strategic value/risk associated with data stored or processed in the cloud? How vulnerable is your own organization to security threats relative to cloud providers?
4. Disaster Recovery: Can the cloud provider meet the RTO/RPO requirements associated with this project/application/data?
5. Performance: What is the cloud provider's SLA commitment and track record in meeting that commitment? Are there resources available for performance monitoring? What impact will latency have on performance?
6. Architecture & Integration: To what extent is the performance of this application/project dependent on integration with other applications or data? Will we have to customize the application to work in a cloud environment?
7. Vendor Support: Does the cloud provider provide migration support, and support for service/performance issues?
8. Vendor Compliance: Does the cloud provider meet all necessary regulatory requirements for this project/application/data? Are there comparable instances supported by the provider meeting the same requirements? Is the provider open to audit by an external agency?
9. Vendor Health: Is the provider financially stable in a competitive market? Will we be compensated for performance shortfalls?

Tonsetic says you can develop a decision framework like this to figure out "when to cloud." Shvetank Shah, executive director of the IT practice at the Corporate Executive Board, adds: Remember that cloud is not a strategy itself, it's tool (like any other tool, such as virtualization) to achieve business goals.

------------
Related:

• Cloud computing: Gauging the benefits, barriers, hype and reality
• Fred Cummins of EDS: Business Expectations for Cloud Computing