CiRBA Cloud Analysis

Cloud computing is being touted as the next “easy” way to save on data center costs. The reality is that companies comparing Cloud options to existing physical environments see reduced costs, but the real benchmark for comparison is the virtualization of existing infrastructure, a comparison that isn’t as straightforward as it may appear.

CiRBA's Data Center Intelligence solution enables organizations to assess Cloud offerings or outsourcing options to determine their fit with an organization's systems, as well as their efficiency and financial objectives. CiRBA's analysis reveals which systems are good candidates to migrate, and determines what allocations and/or hosting tiers are required by examining a number of business, configuration, and utilization considerations.

For example, sensitive data, regulatory requirements, utilization patterns, and utilization personalities all impact whether or not applications are suitable for the Cloud. Once suitable candidates have been identified, a rigorous analysis examines factors, including SLA requirements, costs, operational windows, and utilization personalities and patterns, enabling organizations to fully understand the potential of Cloud alternatives.

Essential to determining whether or not a Cloud option is the best alternative is a comparison against other options, such as virtualizing onto internal servers against hosting in external Clouds (or other outsourcing models). This analysis can yield surprising results, and which option is preferable depends upon many aspects of the analyzed IT environment. These considerations are automatically considered by CiRBA analysis, and include the following:

Application Utilization Levels

Because Cloud environments typically charge for allocated capacity, not actual use, customers pay for unused whitespace.  Cloud environments require an entire instance for each workload, even if only a fraction of its capacity is used.  This can have a dramatic effect when compared to virtualization, where whitespace is pooled and, when analyzed properly, allows many workloads to run on few servers (thereby allowing costs to more closely align with actual use).

Online Windows

Applications that are not required to be online around the clock tend to favor Cloud environments, as instances can be turned off when not required, thereby not incurring charges.  For example, if applications are only required to be available during business hours (M-F 9-5), costs can be reduced significantly, although administrative burdens may increase.

Performance Requirements

Applications that have low performance requirements, such as dev/test or batch conversions, are suitable for Cloud use, as they can be sized in smaller instances, and making better use of the purchased capacity.

External I/O Requirements

Clouds have separate charges for network I/O that go off their local segment, making analysis of I/O rates and patterns essential in determining if a server is a good candidate.

Storage Requirements

Persistent storage volume is also a factor in the cost of Cloud environments, which is typically charged for by volume and duration (i.e. GB months). Therefore, analyzing storage requirements of source servers is also a requirement when analyzing for Clouds.

Sensitive Data

Servers with sensitive data, regulatory requirements, export control restrictions, or other business constraints, are typically not candidates for hosting within Cloud environments. There is currently no mechanism to guarantee that Cloud administrators cannot access sensitive data, and few organizations would trust sensitive customer data to this type of outsourced model.

Availability Requirements

Cloud environments cannot guarantee high levels of availability. For example, Amazon’s EC2 SLA specifies a 99.95% availability level, which may not be suitable for all applications, and excludes failures in the Xen infrastructure, for which they do not take responsibility. CiRBA’s built-in business constraint analysis includes coverage of availability requirements, allowing such cases to be easily uncovered.

These are just some of the considerations when determining whether the Cloud model is suitable for IT systems. CiRBA enables organizations to leverage Amazon EC2 models and customize rules to analyze options according to their own requirements. CiRBA’s detailed What-If analysis of IT environments allows organizations to contrast Cloud models against internal options, such as virtualization, database consolidation, data center consolidation, strategic outsourcing, and the use of internally-hosted utility computing environments, to determine the optimal path for each unique environment.