According to the latest forecast from Gartner, Inc Worldwide, end-user spending on public cloud services is forecast to grow 23.1% in 2021 to total $332.3 billion, up from $270 billion in 2020. While cloud spending is increasing at a rapid pace, most organizations will not make a single leap to the cloud and will inevitably have a mix of both On-premises and Cloud environments. While traditional On-premises environment costs are more easily tracked and managed due to purchasing model (e.g. buy once, upgrade every few years), cloud spend is more dynamic, using a pay as you go model, provisioning/decommissioning resources in days or even hours rather than years changing the way we currently project and manage costs. With this in mind, modernizing cost planning and management is critical for any organization to realize the benefits of a hybrid cloud environment, taking into account both the cost reliability of On-premises assets and the dynamic nature of cloud utilization. There are several methods that can be incorporated from the beginning which can assist with this objective when moving to a hybrid IT environment. FinOps, or Cloud financial management, is one such method.
Forecast Workload Patterns Ahead of Time
The first step in modernizing cost management via FinOps is understanding and identifying forecasted cloud spend. Analysis of customer workloads is an important part of this process and can help determine the most cost-effective pricing model. This includes determining if workloads are intermittent vs. constant. This analysis will provide the most cost-effective blend of services and billing, while still fulfilling mission-critical IT needs. As an example, intermittent workloads can be billed with utilization costs as computing resources are spun up to process workload, and then shut back down when no longer needed using cloud native services such as autoscaling and serverless features, whereas constant workloads may yield the benefit of reserved instances at a lower price point over time. A second aspect of forecasting workload patterns is through performance modeling of processing and throughput to properly size the environment and inform the provisioning of appropriate cloud resources (i.e., instance type). This is done in a phased approach, through modeling the workload upfront followed by subsequent testing to validate sizing and performance need assumptions. Modeling workloads allows organizations to establish a realistic budget that can be more effectively tracked and managed.
Understanding Interfaces and Data Movement
Most public cloud providers charge egress fees to move data out of their cloud and vary by provider and amount of data being moved. Often, those just starting their journey to the cloud may not realize this cost when moving systems and can be surprised by large cost overruns depending on the amount of data being transferred between environments. To avoid these surprises, it is important to understand the system interfaces, frequency and amount of data transfer between the systems On-prem and the various Cloud Service Providers (CSPs). Simply doing a lift and shift of systems without this analysis could result in larger bills than expected and needs to be considered as part of the cloud migration strategy. Another key aspect is the frequency in which data is accessed and identifying the appropriate storage class to minimize fees. Data accessed less frequently can be stored in lower cost storage classes and generate savings over time. Jason Lickel, Vice President of Engineering for T-Rex states, “Understanding where the data resides and where it is moving can significantly reduce overall cloud spend. One mechanism we found effective was the automation of the Interface Control documentation, which identified the data elements and provider and consumers of the data across a large-scale System of Systems, greatly reducing the amount of effort managing the interfaces and provided visibility into the movement of data across the systems prior to migrating to the cloud.”
Monitoring and Alerting
There are several cloud native and third-party billing and cost management tools that can help ensure visibility in near real time in a hybrid cloud environment. Such tools can configure alerts and thresholds set by the organization based on their allocated budget which allows for proactive cost control. In addition, it has been estimated that 30% of cloud resources go unutilized, adding unnecessary cost to organizations. Using CSP features such as accounts and tagging resources to specific environments, systems, and services, organizations can pinpoint potential improvements and cost control actions such as shutting down unused resources or use of right sizing features as described earlier. Tami Curtner, Director of Cloud and Infrastructure Optimization for T-Rex states, “Having visibility and alerting into cloud resource utilization is key to controlling unnecessary cloud spend and drives accountability across the organization. For one agency with an annual cloud consumption of $100M, we established a baseline cloud spend upfront, and utilizing AWS native features, we were able to provide cost breakdowns to system owners monthly, down to the resource level, so funding needs could be planned more effectively.”
Cloud Governance with Flexibility
Ultimately, incorporation of governance is required to establish guardrails and accountability for any agency or organization contemplating a move to a hybrid cloud environment. Cloud governance provides a framework for cloud utilization by defining policies by which cloud resources are provisioned and managed. With the ability to spin up resources through a few clicks of a button, it is imperative the appropriate mechanisms are in place to avoid unplanned costs. Such mechanisms include the incorporation of change control boards whereby any request for additional resources is reviewed by technical and program leadership, against the established baseline, to understand the cost impact and determine if alternate technical solutions are available to meet the requirements. However, flexibility is key here, the governance needs to provide ability to address emergent needs and take advantage of the elasticity of the cloud without being overburdensome such that the process impacts ongoing operations.
As infrastructure modernization continues to advance in IT organizations and Federal agencies, modernization of cost planning and management through methods such as FinOps needs to occur concurrently. T-Rex has proven experience to help navigate the upfront architecture, design, and modeling, implementation of the monitoring and alerts and incorporation of guardrails, preparing agencies to operate efficiently in a hybrid IT environment.
Learn more about T-Rex’s Cloud Adoption & Infrastructure optimization capability.