Cloud Cost Management

Alessandro Mirani

You plan to open a new online business, or you want to revamp your facility by switching to the newest and flamboyant cloud technologies. Your innovative ideas and unstoppable spirit will surely launch you on an exciting adventure, but one that would be short-lived if expenses exceeded revenues. After all, revenue management is not the only challenge for a business; brilliant cost management is the other side of the coin that successful entrepreneurs wear.

New cloud technologies offer several ways to save and optimize on costs, however, it is not impossible to fall into easy pitfalls that have led others before you to have to review not only the composition of their infrastructure but their entire migration strategy.

To best calculate cloud migration costs, you can follow simple but effective guidelines detailed in the following sections.

How to approach costs in a migration strategy.

Starting with the fundamentals, migrating to a public cloud structure could be defined, from a purely accounting perspective, as the transformation of CaPex (Capital Expenditure) costs to OpEx (Operational Expenditure). In other words: expenditures that were previously incurred to make a capital investment (real estate, machinery, licenses, etc.) now become operational expenditures (variable costs associated with increasing/decreasing operations). This, because the vast majority of public cloud services (the totality of the major providers) charge their customers expenses based on usage. From an economic perspective, however, investing in operations, as opposed to capital investment, has advantages and disadvantages.

Investing in capital usually requires you to allocate large sums of money up front. This need may force you to take a portion of the required amount in debt, adding the cost of hypothetical interest to the cost of the principal. The time horizon and guarantees required, in this pattern of expenses, leave little room for flexibility and introduce risks related to the uncertainty inherent in having to incur an expense over time with no way to extricate yourself. Purchase expenses then often call for operating and maintenance expenses. The impressiveness of these decisions, moreover, often requires decision-making bodies and top management to give the go-ahead to proceed with the purchase; a mere formality that no less requires time and energy to execute.

Not only do operating expenses not have these shortcomings, they are deductible expenses, that is, they reduce your company’s taxable income taxes. All these advantages make migrating to the public cloud particularly attractive, but all that glitters is not gold. First, not owning infrastructure that is essential to your operations may not only be a strategic problem, but also a legal one. Also, you decide how to write off the cost of a capital purchase, and barring mistakes, you can calculate without major surprises how much it will cost year to year to, say, buy a server today. In an operating cost model, it is supply that decides year by year costs, and as demand increases, supply increases costs. Having to rely on market stability, is certainly a risk.

By being clear about these differences you will know how best to evaluate the expenses those who migrate to public clouds incur. How exactly to calculate these expenses, however? Without further ado, the answer is in the following section.

Most common sources of cost and how to control them

In 2019, research by claimed that more than 50 percent of companies that had migrated to the cloud had run into overspending compared to initial expectations. It is therefore essential not to underestimate this aspect during the migration and to know from the outset which errors have the most impact when calculating the cost of migrating to a public cloud:

Secondary costs: many storage services (data storage), are a simple services that offer storage space for a few cents per gigabyte/month. However, these services, especially in their cheaper versions, also charge a fee per access. Simple data access costs ten times as much as storage per gigabyte. This example shows how, for some costs, the primary service is tied to higher secondary costs than the primary service.

Geolocation: having the same service in multiple locations allows us to distribute risk and improves our disaster recovery capability. Location, however, often affects the cost of a service offered. Two virtual machines that have the same memory, the same processor, and run in the same time frame will have different costs if they are located in two different locations. This logic applies, in some cases, even within the same country. The United States, for example, is a region that often includes several areas, offering the same services but at a significantly different cost.

Large-scale costs: transferring data between data centers, between servers, or between virtual networks not physically in the same area is a service often required by cloud infrastructure users. Imagine having to transfer a small file between ten locations. If by chance the file size were to suddenly change, the cost impact would be harmless…unless that file is transferred thousands of times a day. Failure to consider that some operations are repeated hundreds of times, in a small time frame, leads one to underestimate the potential impact of a contingency related to these expenses.

The fact that small numbers on the “menu” translate into large numbers once you go the cloud migration route is therefore no mystery. This known problem, however, has solutions already available that you can employ right away. The major public cloud providers all offer a specific pricing tool (price calculator), which allows you to make accurate forecasts of expenses. As a preventive measure, some effective techniques for containing costs are:

Resource planning: plan ahead for the times and amounts of resources you want to deploy to increase the likelihood of staying within budgets. Setting a high threshold for utilization of deployed infrastructure or shutting down, if not decommissioning, unused applications are the main ways to avoid breaking through allocated budgets.

Set usage limits: many cloud applications allow you to set a resource usage threshold, or even an actual money limit, at which operations are automatically stopped.

Use bonuses and promotions: discounts, concessions and promotions are available on almost every platform and can help you save a large percentage of your spending. To access these discounts, you often only need to change certain conditions of using a service (only within certain hours, prepayment).

This is as far as periodic costs associated with operations are concerned. To migrate to the cloud, however, you will incur other expenses: consulting, developments, data transfer and integrations are all cost items that you will have to take into account in the early stages of migration. Should everything be right with this? Not exactly.

After migrating to the public cloud?

What will you have to pay upon completion of the migration to the cloud? In addition to periodic costs such as those listed above, you have to take into account other items such as staff training, monitoring compliance against the standards you need, and infrastructure management and coordination. In other words, all expenses that on your side might vary simply because, now, the composition of your infrastructure has varied.

Of course, these costs do not necessarily vary to your disadvantage, but what I want to suggest is to take a holistic and realistic approach to estimating expenses. Small distractions can translate into big problems the moment you are not in total control of your expenses. However, this does not detract from the benefits of migrating to the cloud. Greater flexibility, infrastructure that is always up-to-date and on which you are not the one doing the maintenance can be a huge weight in the balance of your decisions.

The cost transparency that the major public cloud providers offer is a benefit I would like to emphasize. Whatever cost structure you choose, a reliable and clear provider is a key element in saving time and resources.

Leave a Comment