CLOUD COMPUTING: MYTHS AND REALITIES(2)

By EMEKA AGINAM
Claim 5: The Cloud Will Always Save Money on IT
Cost reduction is a primary business motivator for adopting cloud computing, but quantifying cost saving in building a business case is not always easy.

While quantifying cost savings is relatively easy for SaaS but becomes increasingly more difficult when it comes to PaaS and IaaS.

PaaS and Iaas often have factors that are difficult to measure, such as application modification time, running pilots, and the cost of training a staff in cloud computing skills, amongst many others.

The cloud offers financial savings in more than one way, however. The most obvious is the potential to reduce IT spending on existing requirements.

Another is how it allows a business to be more agile and increases revenue by providing low-cost, low-risk self-provisioning IT. This allows organisations to explore new opportunities that may have previously presented too high an investment risk/reward ratio.

In summary, when it comes to saving money on IT, small- and medium-size organisations can generally benefit from leveraging cloud computing. Larger organisations may not benefit financially from cloud computing due to the economies of scale that they can apply to on-premises IT provisioning. However, other advantages such as on-demand provisioning and reduced administration cloud computing may prove beneficial to the business and offset the cost differences.

Claim 6: Cloud Computing Resolves Performance Problems

Elasticity and a seemingly infinite amount of compute and storage capacity are often touted as major benefits of the cloud, but while these are clearly significant benefits, they are not measures of application performance. Poor performance issues are often caused by infrastructure capacity problems rather than poor application implementation. If an application running on a server becomes overloaded and all users are affected, then the users suffer poor response times, plus the application has a performance problem.

Cloud computing can certainly provide solutions to infrastructure capacity problems due to its elastic scalability, but performance will often vary. PaaS delivery models such as Google App Engine and Microsoft Azure provide satisfactory performance. The Google App Engine for example will provide an application with all the compute and storage capacity required, but it will not be high performance. Only IaaS can provide high performance, and even this isn't always a sure bet.

There are a number of reasons why this is so. Cloud providers may oversell their capacity, causing performance problems that are totally out of the hands of the cloud user. In addition, virtualisation software runs on a multitenant architecture, and what other tenants are undertaking on that infrastructure plays a significant part in performance at any particular moment in time. So while the cloud provides the ability to improve performance problems through increased capacity, only satisfactory or acceptable performance, rather than high performance, should be expected.

Claim 7: All Software Should Be Moved to the Cloud
The attractive benefits of the cloud may encourage organisations to try to move all computing to the cloud. For some organisations, reducing the on-premises IT to an absolute minimum is appropriate. For others, internal IT departments with software development teams are essential as the bespoke software they develop enables the organisation to gain competitive advantage.

To be continued