Cloud infrastructure Accounting (part 1)

The whole idea of XaaS is the ability of the customers to add or subtract users or computing power (thus, workload on the vendor’s infrastructure) as they please, because they have tangible monetary benefit from this fluctuation; they pay for more only when they really need it and when they don’t their monthly expenses go down.

This fluctuation creates spikes and valleys in the infrastructure of the vendor, which follow the customers’ “horsepower” needs. There are other metrics, though, that are never (or almost never) decreased. Perhaps the rate of increase fluctuates but it is almost never negative. The most characteristic example is this of the consumed disk space in the infrastructure: it is always increasing (unless of course a massive number of customers leave the service, which is highly unlikely).

These considerations create the need for some kind of accounting tracking of the infrastructure that is being used (or not used) and also some kind of depiction in financial statementsthat possible venture capitalists or other investors may require from the vendor’s management. In this and the next post I shall deal with this issue of accounting for used/unused XaaS capacity.

The XaaS infrastructure consists of several tangible metrics which are paid for and tracked by accounting. Typically, a XaaS service has the following three measures:

Assets such as hardware, servers, routers etc. which are purchased and tracked like fixed assets. These assets are depreciated, monthly or annually, according to local accounting practices.

Expenses such as payroll costs for the infrastructure personnel which keeps it running (operators, technicians etc.). Also, other expenses to third parties, such as system software licenses, service contracts with the hardware vendors, consulting fees etc.

Income which is generated by the sales of the XaaS service to customers.

The above three main axes will give you the monetary result of the XaaS service: whether it is profitable, how much does it earn etc. But what about non-tangible and non-monetary values that are part of the service? They are not tracked by the above accounting system. Here are some examples:

How much disk space has the service? How much of this is already occupied by customer data? Do we need more? How much more?

How many processor cores has the service? Are they being used above 60% or have we done the mistake of purchasing more than we need?

How many virtual machines have we set up in the infrastructure and how many of those are being used? What is the average usage of each?

And on the time scale: How much time does it take us to manage all those servers? How much time do we consume, monthly for the upgrades and antivirus protection of them? Do we need more personnel to do the task? How much more?

Of course, if the infrastructure is “rented” to a third party IaaS/PaaS provider on behalf of the SaaS vendor, perhaps some of the above questions are not valid. But let’s assume for argument’s sake that one vendor provides both the software (SaaS) and the infrastructure to run on (IaaS or PaaS). Accounting knows how much money they have paid for this infrastructure to come to life. But does the management or the COO know how much of it is being used? When the CTO comes asking for “more power”, is he right or is he wrong? And, bottom line, if the SaaS customer enjoys lower Operating Cost due to their decision to adopt SaaS, shouldn’t the SaaS vendor enjoy similar economies of scale? Of course, yes!

In the following post we shall see a practical proposition of how such intangible or non-monetary metrics can be tracked and accounted for and also how can their history be recorded and tracked. Stay tuned!