Cloud infrastructure Accounting (part 2)

In the previous post we made a brief reference to “tangible” costs and income that the cloud infrastructure incurs/generates to the XaaS vendor. Such issues are well known and adequately followed by standard accounting practices (charging expenses, recording the purchase of assets and tracking their depreciation and, of course, customer invoicing for your income).

The new dimension that we discussed is this of “intangible” or “non-monetary” values that exist in a XaaS infrastructure, and especially the magnitude of unused capacity that is there. In this post I will demonstrate that basic accounting rules and handling can help keep track of these non-monetary values in a way that is transparent and, hopefully, acceptable by the majority of finance managers.

First you need to define what these values are: I spoke about disk space (used and unused), number of physical machines, logical machines and processor cores and average processor usage. Other dimensions may also exist… I also spoke about the time dimension: how long does it take you to perform infrastructure maintenance stuff (e.g. database administration etc.).

After you have defined these values or “dimensions” you can setup a basic accounting structure to keep track of them, like this:

Create an account for each of these dimensions. Typically these should be memo accounts (not like expenses and income; P&L and Balance Sheet accounts). So, you will have accounts like “DISK_SPACE”, “NUMBER_OF_VIRTUAL_MACHINES” etc.

Each time you add something to your infrastructure you may debit the respective account. E.g. if you added 100GB of new disk space you can debit the respective account with the “amount” of “100,000,000”. If you added 5 new VM’s you should debit the VM account with “5”, etc.

Each time you “consume” something of the above, you must post a credit to the respective account. E.g. if 10MB of disk space have been consumed since the last checking (what is the last checking? see below) you should credit with “10,000” etc.

Following the above logic, the balance of each account will always give you the unused size or number of the respective metric. Furthermore, using standard accounting reports such as Trial Balance, General Ledger etc. you can track the evolution of each metric. How much did you consume this month, when the peak period of consumption during the year etc. was are just a few of the questions that you can answer…

A very important issue in order to make this system workable is this: although the “debits” are quite easy to follow (each time you add something to you infrastructure, you just make a debit, whose real-life equivalent is the vendor’s invoice for the new server, the new router some payroll cost for the administrators jobs execution etc.), the credits need to be tracked following a specific internal process: disk space consumption happens every day (every second, in fact!) and keeping track of it in real-time is impossible. Therefore, you need to establish procedures for the periodic tracking and reporting. For example, you may have your technical manager prepare monthly reports on “consumption” of disk space, processor usage etc. and you translate them into memo accounting entries.

Having surpassed these problems, you now have an accounting system in place to track your resources consumption. If you want to go two steps further with that, here are a couple of more ideas:

Some of the metrics that we have discussed may be possible (and business-sensible) to keep track per customer. For example, disk space consumed per customer may be useful (or even a must, if you have a GB-consumption-based-SLA!), while number of processor cores is irrelevant. In these cases you may want to develop those memo accounts to subaccounts that represent your customers. That way, instead of having “DISK_SPACE”, you would work with:
“DISK_SPACE_CUSTOMER_A”,
“DISK_SPACE_CUSTOMER_B”
etc.
Or better yet, you might want to consider using “cost center accounting” for those accounts. So, the unique “DISK_SPACE” account is analyzed to multiple cost centers that represent your customers (hmmm... cost centers in memo accounts… I wonder what a true accountant has to say about this!)

If you have an accounting system keeping track of “real” events, it is most probable that you can also devise a budgeting system to foresee or budget future needs for these sizes: You finance manager is doing it already: He sees the expenses of the current year and budgets for next year. In the same manner, you could be looking at disk space consumption and customer number increase tendency and budget for next year’s needs in disk space (or number of servers, processor cores etc.)

By mixing and matching technical knowledge and basic accounting principles you can establish procedures and metrics to keep track of you XaaS infrastructure and better plan for the future. It will also help you identify areas of under-consumption and probable opportunities for cost cutting. While this is in the daily schedule of your customers, now you can also deal with it.