Chargeback for cloud services can be a controversial subject—not necessarily because business units don’t want to be charged for the cloud services they consume, but because the factors that impact chargeback for on-premises IT services may not be acceptable or effective in a self-provisioning cloud environment.
Implementing an acceptable and effective system of chargeback for cloud services can have multiple benefits for a company. It makes business units aware of the resources they are consuming (often resulting in a reduction of consumption) and encourages business managers to make better decisions about their resource requirements and priorities, improving the performance of the whole company.
In order to be acceptable, a system of chargeback for cloud services has to be fair. Historically, in pre-cloud chargeback models, fairness was determined by four factors—transparency, understandability, accuracy, and controllability to the extent a business unit could control its charged-back IT costs by controlling its consumption of IT services. Low controllability often led to resentment of the system.
In a self-provisioning cloud environment, these four factors are sometimes not enough to be either acceptable or effective. With a variable pay-as-you-go model, resources can be made available or scaled within a few clicks. This makes it challenging to accurately track and predict OpEx spend in the cloud. While a cloud model shifts most spend to the pay as you go model, some cloud services can be committed to and paid for upfront adding another layer of complexity. Whereas within an on-premises IT infrastructure, charged-back IT costs are fairly standard, business managers in a cloud environment need to know if their IT costs are going to increase above budget; and, if so, they need to know where costs are being driven up, and what for.
“How to do Successful Chargeback for Cloud Services”
In December 2011, a paper was presented to the Conference on the Economics of Grids, Clouds, Systems, and Services (GECON) entitled “How to do Successful Chargeback for Cloud Services” (PDF). The authors from Utrecht University addressed the shortcomings of historic chargeback systems by introducing four additional factors—predictability, measurability, accountability, and comparability.
Three of the factors (predictability, measurability and accountability) were included to enhance the acceptability of chargeback systems. During their research, the authors found that one of business unit managers’ main concerns about the cloud was receiving higher-than-expected chargeback invoices. The factors of measurability and accountability are intended to eliminate disputes over chargeback amounts.
The comparability factor is interesting because it was intended to give business unit managers the option of comparing the prices of internally provisioned IT services against equivalent services available on the open market. The objective being to increase price consciousness and maximize cost efficiencies. In hindsight, this isn’t such a great idea because it can lead to the development of Shadow IT.
At the time the paper was presented, Shadow IT was already an issue. Indeed, also in December 2011, Gartner predicted 35 percent of IT expenditure would be managed outside the IT department by 2015. However, cost isn’t the only concern for companies with a Shadow IT problem, as the uncontrolled use of unapproved resources can also lead to inefficiency, security gaps, and compliance issues.
Making a Cloud Chargeback System Acceptable and Effective
The way to make a cloud chargeback system acceptable and effective (and overcome any potential Shadow IT issues) is through tagging. User-defined tags were first introduced by Amazon Web Services (AWS) for EC2 resources in 2010 in order to simplify the administration of cloud infrastructures, and now many more resource beyond compute resources can be tagged on AWS, Azure, and Google Cloud.
Tagging enables companies to allocate cloud costs to business units in a transparent, understandable, accurate, and controllable manner. Depending on which Cloud Service Provider is used, companies can allocate up to sixty-four tags per resource, allowing for a degree of granularity that will satisfy any business unit manager’s desire for measurability and accountability.
Business unit managers will be able to draw down reports at any stage of the month to ensure their unit´s cloud costs are within budget, and optimize their resource usage where necessary. Finally, although tagging won’t physically prevent the development of a Shadow IT environment, any cloud costs that cannot be accounted for should be relatively simple to track down.
To make sure that every unit within the business accepts the cloud chargeback system, it’s a good idea to trial the system in showback mode first—i.e. not necessarily charging each unit for its usage of IT services, but explaining to units where costs have accrued in order to facilitate the optimization of resources and a change in consumption behavior before the cloud chargeback system goes live.