For today’s CFOs, IT is at the top of their agenda. With 26 percent of IT investments requiring direct authorization of a CFO in an organization, several CFOs have either embraced or are ready to adopt cloud due its OPEX model. To this end, Gartner estimates that by 2020, the aggregate amount of cloud shift might reach up to $216B. And with AWS topping the charts among the CSPs, Wikibon estimates that by 2022 AWS will reach $43B in revenue, and will be 8.2% of all cloud spending. Despite exponential increase in the adoption, there is one major fear attached to AWS, for that matter all the cloud’s adoption — how to be on top of cloud sprawl, and how to perfect AWS cost forecasting and budgeting as an enterprise business.
If you are a CFO trying to up your game and seeking to build a roadmap for AWS cloud cost modelling, spend forecasting and cloud budgeting, and above all assuage cloud sprawl? Here’s how:
AWS Cost Modeling and Calculating TCO of the Cloud
Unlike owned and self-operated data centers, cost modeling for procuring an instances over a specific cloud is different. Hence, total cost of ownership (TCO) is different in both the cases. When it comes to calculating the TCO of your cloud infrastructure, the real challenge is figuring out the pricing models of cloud service providers and planning the accurate capacity requirements for a period of at least two quarters to one full year. And when it comes to AWS cloud, all the services need to be taken into account including data, storage, networking on top of compute infrastructure, etc.
As a CFO, you also need an application or workload specific cost modeling on AWS cloud. In this case, the first step should be to take business demand variations into account as per periodic seasonal cycles. Your cost estimations will definitely take a hit if you do not consider the impact of the seasonal fluctuations on service-level agreement, usage patterns, storage requirements, and cloud environment configurations.
Plus, you need to detail the capacity plan to the level of individual application workloads instead of other dimensions like departments. Because the impact will vary from workload to workload. There are chances, you might be over planning or under planning the capacities.
While centralized planning at the application or workload level for all departments is unrealistic and hard to do, it can be an apt approach for realistic budgeting and forecasting if it’s already followed for data centers. However, when departments or business units are asked to create cloud cost forecasting and budgeting plans, you can indirectly work with them to create the cost plan at the application workload level.
Leverage Cloud Cost Reports to Identify Peak Resource Usage Scenarios
Identifying peak resource utilization patterns for various workloads is one of the stepping stones to a reliable AWS cost budgeting and forecasting. This can be achieved easily through periodic analytics and running reports over the usage data. The best way forward is to leverage other data sources like seasonal customer demand patterns and see whether these patterns have correlations with the peak resource usages. When these patterns are identified in advance, required base capacity can be taken care by buying pertinent reserved instances (RIs).
However, use of RIs requires proportional amortization allocation that are based on usage hours across specific business units or application workloads. It may also involve upfront capital expenditure (CAPEX) depending on the cloud provider’s business models.
On the other hand, even though RIs take care of predictable usage requirements, they cannot help with spikes. Plus, amortizing RI costs to the appropriate projects or teams (via resource tags) within the organization is a huge challenge too.
Planning Cloud Capacity and RI Amortization
A proportional cloud capacity purchase and amortization allocation based on usage hours requires several measures, like mapping a piece of the upfront RI payment to each hour of usage for the RI, using the workload metadata tags that can be assigned to resources and tieing back to RI usage ( so that proper cost allocation can be done at cost center level or at business unit level). By doing so, you have visibility into expected increase in cost by various cost centers or business units within the organization.
If you want to know the proportional and hourly amortized costs of RI, few smart third-party tools can throw usage insights that reflect cost incurred by various business units or teams’ RI spend across your organization. The next step is to track overall cloud as well as workload-level consumption against the planned budget so that as a business you can be within the allocated expenditure for cloud usage.
Track Cloud Consumption Against the Plan through Cloud Management Platforms
The organizational level cloud budgeting and cost plan adopted must be regularly tracked at monthly or bi-monthly intervals by comparing the planned usage vs. actual cloud spend so that right financial governance controls can be deployed.
Also a periodic review of planned vs actual usage will help you and your business to understand how to optimize the use of underutilized cloud resources across organization or within specific cost centers to reduce AWS cloud expenditure.
Using tools like Botmetric, you can identify the anomalous usage patterns, detect underused resources and take corrective action in real-time instead of waiting till the end of the quarter period to take stock of overall forecasting.
Another key aspect that cannot be ignored by financial controllers handling the cloud procurement within the organization is to streamline AWS cloud cost budgeting as per organizational level chargeback or internal invoicing based on business units or cost centers. The best way forward to achieve this goal is to plan minimal governance around resource costs at department level or workload level through tagging, allocating extra spend to specific teams and generating chargeback invoices to respective business units.
Applying Predictive Analytics for Projecting Cloud Spend and Forecasting Usage
Even though AWS itself provides a basic AWS Cost Explorer Forecast widget in the AWS console, you will need a tool beyond that, which can help achieveforecasting with ease. You can leverage predictive data analytics tools like Botmetric to keep your cloud spend in control by acting upon smart, intelligent recommendations based on more than 80 AWS best practices.
Also, Botmetric provides predictive Analytics as a way to get forecasting data, deploying financial controls for organizational level governance and meet future demand patterns accurately based on the past usage data.
Adopt Financial Governance Best Practices for Cost Management
While the above four approaches help you at the macro-level-cost-management level within a business and at the cost center level, the best way to optimize your cloud usage and spend is by adopting the best practices advocated by the cloud experts in the ecosystem, such as:
- Track storage and data transfer charges at cost center or business unit level
- Constantly monitor and remove unused resources at workload level
- Track, manage, and fix underused instances at workload level
- Look for unattached persistent volumes and old snapshots at organizational level
- Look for allocated spend spikes to specific business units or cost centers through chargebacks
There are many more best practices. However, to follow all the best practices, decreed by AWS as well as cloud experts, requires a lot of effort and spending on resources. Nevertheless, there is a way out — using intelligent, AI-powered platforms.
To succeed as a CFO in cloud world, access to consolidated billing report, visibility into cloud data to monitor usage, capability to establish governance policies to control cloud expenses, and successfully forecast future expenses are a must haves. And with access to tools that can provide real-time notifications of exceeded budget limits, budget forecasting, and visibility into your cloud infrastructure costs will be like icing on the cake. Botmetric Cost & Governance provides just that.
With Botmetric you can spend more time determining the best way to optimize your cloud with your team rather than analyzing complex billing reports, following trends on specific infrastructure growth, and continuously monitoring on-going assessments of your cloud spend. Give it a try !