What does an application in my datacenter cost, and what are the financial consequences of going to the cloud? Seemingly not a very complicated question. Certainly an organization with a cloud strategy should be able to answer this. For that reason the financial analysis, a.k.a. the business case, is an important part of a cloud strategy. But how detailed and exact does your analysis need to be and what will you do with all that insight?
What are the financial consequences of going to the cloud?
The purpose of a financial analysis is to provide insight into the financial consequences of the various cloud movements. In other words, what happens when I pick up this application and bring it to the cloud. Every organization has its own motives for bringing their IT landscape to the cloud. Getting a grip on costs and optimizing expenditure is an important motive (if somewhat short-sighted). We want our organization to select the optimum landing site per application, to do that we need to include the business case in our considerations and for that insight into the current costs per application is required.
This insight into usage costs requires a financial model with which multiple ‘views’ of the costs can be made. For example, you want to be able to make a ‘cross-section’ of all the costs associated with one application or a cross-section of the costs incurred for one type of platform (for example: what does it cost to maintain the Linux servers in our own data center).
How deep you have to dig for data depends on the application. In general, the more detailed and reliable the information is that is fed into the model the more granular you can investigate and analyze different scenarios. Sometimes it is difficult to make connections between a multitude of sources of information, this requires; translation, interpretation and making assumptions (validated ones of course).
Examples of data sources for financial analysis:
- In order to compare the costs of a server in the data center with the costs of cloud, you do not only need all the investments and maintenance costs of the data center, but you also need to know what net capacity the data center provides.
- To determine what an application costs you have to, among other things, know the infrastructure footprint of an application (how many servers, storage, memory, licenses, etc.).
- To accurately allocate the costs of software development, you need to know how many hours a developer spends on a certain application.
Fortunately, it is not necessarily necessary to count per server how many hard disks it contains or ask employees to estimate how many hours are spent on the management of a certain application. In many cases it is sufficient to work with standard units like the average capacity of a server. There is also plenty benchmark data available; information about what the costs are like for comparable organizations. You can then use that benchmark data to supplement any missing data.
How to segment the financial model
As mentioned, the required level of detail depends on the purpose of the analysis. For some infrastructure-oriented organizations, it is necessary to dig all the way to the level of storage and memory capacity. Only then can you compare apples with apples in the different cloud models. If the strategy mainly involves an increase in the uptake of SaaS applications this level of detail is too much; on the other hand, you may want to pay more attention to the costs of for instance technical application management and software development.
To keep things clear, it is wise to consolidate the costs into layers and only zoom in when necessary.
A straight forward approach to the layers you want to consolidate the information into.
- Infrastructure (incl. Management)
- Software (purchasing licenses and maintenance)
- Application, middleware and database management (both technical and functional management)
- Software development (customization, extensions, maintenance, etc.)
Using layers to segment the data allows you to distinguish between, for example, the linux and the Oracle platform or between high-performance storage and archive storage.
How to build the business case
In order to develop this analysis into a business case, an extra level must be added; investments (transition costs), the new operating costs and the expected revenues. In other words; how much do I have to invest in the transition and what are my costs after the transition. Only then can you determine if a cloud migration is financially interesting in the short and longer term. In addition, the cloud strategy may contain objectives that cannot be quantified immediately but are a major driver for the transition. Use weights in your financial model to represent these drivers.
For IaaS and PaaS the new costs are reasonably easy to calculate with the help of the online calculators such as the ones AWS and Azure offer. The transition costs (costs of the migration) are a lot more difficult to calculate, the bandwidth estimate lies between 2,000 and 3,500 per server to be migrated.
Keep an eye on the purpose of the analysis when analyzing; it is very tempting to analyze ever more deeply. Avoid the situation where analysis leads to delay.