With many pilot cloud projects gathering steam, organizations are evaluating transitioning their IT systems to a cloud-based architecture. However, such a full-scale move must take into account security risks, lock-in risks, and cost-benefit analysis over the lifetime of the investment. An InformationWeek Analytics report outlined a comprehensive survey of 393 individuals within various companies, 28% of whom had more than 10,000 employees. Amongst the many findings within the report, the most interesting ones were:
- 34% of respondents involved in the cloud used it for SaaS (applications delivered via the cloud), 21% for IaaS (storage or virtual servers delivered via the cloud), 16% for PaaS (web platform delivered via the cloud), and 16% for DaaS (Data-as-a-Service for BI and other data lookup services delivered via the cloud)
- 29% were not using the cloud at all
- More than one-third claimed to build in 31% or more excess server and storage capacity for non-cloud computing systems
- 73% cited "Integration with Enterprise applications" and 69% cited "Cost of Hardware and Software" as factors when choosing a business technology
- Almost 92% exhibited some sort of likelihood to comprehensively carry out an extensive ROI analysis of the expected lifespan of a cloud computing project
- 46% said that their ROI calculation would span 3-5 years
- 45% stated that "elasticity" is frequently or often required
There was also a feeling amongst respondents that cloud computing works for commodity applications but that complex integration requirements make costs skyrocket. The major sources of cost savings touted by cloud proponents involved three areas: efficiencies as a result of economies of scale, use of commodity gear and elasticity.
This last area of elasticity is worth exploring further, especially in light of the number of respondents claiming to require excess capacity for their non-cloud computing applications, and the assertion that complex integration requirements are increasing the costs in the cloud. Elasticity refers to the ability to scale up or scale down on storage resources on the fly. But the reality of elasticity "on-demand" is that most major software vendors don't provide the ability to add CPUs without additional costs, and coding applications that appropriately scale up are difficult. Thus, given the above data about the necessity of elasticity and the large percentage of companies that conduct detailed cloud ROI analysis, it is evident that these two factors are correlated.
This last area of elasticity is worth exploring further, especially in light of the number of respondents claiming to require excess capacity for their non-cloud computing applications, and the assertion that complex integration requirements are increasing the costs in the cloud. Elasticity refers to the ability to scale up or scale down on storage resources on the fly. But the reality of elasticity "on-demand" is that most major software vendors don't provide the ability to add CPUs without additional costs, and coding applications that appropriately scale up are difficult. Thus, given the above data about the necessity of elasticity and the large percentage of companies that conduct detailed cloud ROI analysis, it is evident that these two factors are correlated.
- Increasing the ROI of a Cloud Deployment
- In order for CIOs to see more of an ROI from deploying applications in the cloud, several things must happen:
- Data center automation software must reach a level of sophistication where they are able to automatically coordinate tasks between on-premise and cloud applications to optimize elasticity
- Software vendors must allow for special pricing for cloud providers so that these savings can be passed onto customers - a way to allow this is to ensure a multi-tenant architecture consisting of customers that use the same on-premise software as is used in the cloud-based edition by the respective provider.
- Ensure that the web platform used for PaaS purposes by the customer is compatible with the SaaS applications that they subscribe to, in order to enable any custom widgets that may need to be written
- Massive improvements in the "converged fabric" architecture that brings together servers, storage, and networking, so that pools of additional capacity are easily available where elasticity is needed.
