Cloudware Economics - Guide to Cloud Computing for Business and Technology Managers - page 285

costs, as well as cost of skills upgrade (offset by perhaps a leaner IT
team), that need also be accounted for. By definition, cloud implies a
dynamic service that assures optimal utilization. This, on the other
hand, means that fluctuations in service use could become challeng-
ing, unlike the static resources in a traditional IT environment that
can be accounted for more easily.
2. Incorporate hidden cost. Whereas in-house provisioning incurs hid-
den costs such as additional administrative headcount, additional
property and facilities requirements, inevitable overprovisioning
costs, and additional costs for ensuring redundancy, the cloud pro-
vision's hidden costs could come from potential costs such as ser-
vice interruptions, inappropriate service scaling, mismanagement
or a denial of service attack, extra security, and contingency disaster
preparedness and recovery plan costs, as well as the initial cost of
cloud readiness including costs associated with setup, interfacing
and integrating with discrete local infrastructure or resources, and
administrating the whole new operating system.
3. Evaluate the application profiles and service mix. Applications uti-
lize computing resources at varying rates. Some are more compute
intensive, whereas others do a small amount of processing across an
enormous amount of data. This exercise helps to create a clearer TCO
picture by assigning costs to the different cloud services, according
to application profile.
4. Calculate the TCO under a number of different application topolo-
gies to understand costs under different loads. Identify and cost
the required compute instances according to application load
variations. Technically speaking, study the horizontal and vertical
scaling patterns of the applications. If the load on an application
varies significantly, it will most likely require a larger deploy-
ment of multiple compute instances to reduce the application
bottlenecks.
5. Evaluate the role of load variation. It is important to identify the
periods and patterns of application requiring larger loads or expe-
rience load variation. A static pattern assumption is hardly useful
to calculate cloud TCO. Carrying out a statistical (e.g., Monte Carlo)
and scenario analysis to explicitly assess TCO under different load
patterns can assist with more accurate estimation of TCO.
In practice to derive value from TCO analysis, it should be included in the
calculation of other measures such as return on investment (ROI), net present
value (NPV), internal rate of return (IRR), or Economic Value Added (EVA).
That way, value planning for cloud is not one-dimensionally cost focused,
but it will take into account the quantified business benefits as well.