Determining the true cloud impact and the preparation for cloud implementation must address the application level.
Corporate Facilities and IT departments collectively struggle with the development of Data Center Strategy as it must straddle the competing requirements and perspectives of these two worlds. Facilities is measured on the stability and costs of the data centers. IT conversely, wants no downtime and wants the data centers to always have capacity for their changing, and often unplanned, IT requirements. The overall Corporate strategy objective must be to identify and size the data center mix such that all requirements can be met over time at the lowest cost while also supporting the changing needs of the mission and application portfolio
The cloud’s significance within overall data center strategy and impact on total costs has grown such that any conclusions without considering both the internal savings vs net additional costs from cloud adoption would be missing a big part of the picture.
The impact of cloud adoption is the newest variable in the equation and has become the elephant in the data center strategy room. Indeed, the migration of applications to the cloud has become the biggest driver for change in the enterprise data center environment.
One hears occasional declarations that “everything” is moving to the cloud. Such bold statements quickly fade upon deeper analysis but the impact of application migration to the cloud should not be lost. The cloud’s significance within overall data center strategy and impact on total costs has grown such that any conclusions without considering both the internal savings vs net additional costs from cloud adoption would be missing a big part of the picture. Further, understanding how the cloud economics differ for each application/category should not be sacrificed under the weight of the cloud’s surrounding hype.
Most on-line TCO cost comparison tools fail to make this distinction between fixed and variable costs when they calculate “savings” from the cloud.
Some companies may initially opt for high level estimations regarding cost savings from moving candidate environments (e.g. dev/test) or candidate applications (e.g. web apps) to the cloud rather than a detailed analysis. This form of high-level review may provide directional information regarding whether to include the cloud in the overall strategy and an idea regarding potential costs or savings, but it will not indicate how to proceed or provide the precise information needed to begin migration planning of applications to the cloud. This high-level approach makes sense when generally assessing cloud among multiple data center sourcing options.
Eventually, however, determining the true cloud impact and the preparation for cloud implementation must address the application level and include the following steps:
• Evaluate and categorize candidate applications for cloud readiness based on multiple elements including OS/version, application dependencies, bulkiness/chattiness, latency issues, security factors, licensing etc. Applications can be categorized into “buckets” representing their readiness based on suitability and transformation costs. Based on this categorization, application can be identified as candidates for migration to the cloud over a forecast period.
• Identify the cost savings associated with those applications moving to the cloud. It is important understand the marginal cost savings for each element. That is, the costs that really go away when an application moves to the cloud.
Each underlying application element must be identified, costed, and analyzed to understand what and when costs are saved. These include IT and Facilities cost elements such as virtual machine licenses, server costs, storage (all required types), network, power, and data center costs. The latter represent costs that do not go away linearly as apps and servers move to the cloud. If you move 50 VMs to the cloud, you may save license costs and perhaps some power, but data center rent, depreciation, and maintenance will likely remain unaffected. Most on-line TCO cost comparison tools fail to make this distinction between fixed and variable costs when they calculate “savings” from the cloud.
• Identify the corresponding cost increase to house and run the applications in the cloud. Most cloud providers have intuitive and detailed cost modeling tools necessary to estimate this side of the equation. What they lack, however, is the ability to project multiple applications moving to the cloud over time and how they are growing or changing over time. The cloud provider costing tools need to be augmented with analysis that reflects how your applications are expected to grow over time.
It is important to also add three cost elements that may be beyond the scope of many cloud cost calculators.
Application transformation costs: These include the development costs to adjust, upgrade, or replatform the application to enable it to function in the cloud.
Migration costs: Moving applications to the cloud entails a full migration effort little different from a migration between enterprise data centers. The costs of planning and execution need to be included in the cloud side of the cost equation.
Governance and Management: Depending on the scope of the cloud initiative, it may be necessary to construct a management and governance organization that enables the users to access cloud resources and enforce limits, rules, and standards.
• Complete the longer-term Total Cost Analysis. Complete a year by year cost analysis that compares year by year savings from fewer VMs, lower power, reduced headcount, etc. against the cloud server, IO, storage, transformation, and migration costs. Additionally, after this analysis, the lower data center footprints required can be built into the data center strategy and likely yield lower overall data center outlays.
These steps, while requiring substantial effort, will enable you to quantify the effect of each application’s move to the cloud while also providing much of the early groundwork needed for the actual migration.