Getty Images/iStockphoto
Rein in services to avoid wasted cloud spend
Organizations often make the easy mistake of duplicate purchases, which lead to wasted cloud spend. Learn strategies to avoid overspending in the cloud.
Reining in cloud services is a team effort that requires careful study, analysis and proactivity to ensure organizations reach their cloud cost optimization goals.
Large companies with multiple departments and budgets often make duplicate purchases for the same or similar service. Even seasoned cloud organizations can overestimate the services required to support a specific business goal. But services that aren't used can still cost money.
Estimating cloud spending is a talent as much as it's a skill. Cloud professionals can develop and implement best practices to avoid cloud overspending and overservicing. These best practices start with understanding where cloud overservicing frequently occurs.
Anatomy of cloud overservicing
Many factors can cause enterprises to fall into the trap of cloud overservicing. Here are some of the most common pain points.
Cloud adoption management
Enterprise moves to the cloud are fluctuating, especially in the current corporate spending environment. The rapid, unmanaged growth of cloud adoption inside an organization leads to the inability to keep up with demand. This leads to the overprovisioning of cloud services to ensure performance and availability.
Inversely, cloud repatriation isn't a one-size-fits-all cloud cost-saving strategy. Depending on what data and applications an organization repatriates from the cloud, it might lose the economies of scale that come with moving to the cloud. That decision could result in higher costs for resources the organization keeps in the cloud. In fact, cloud repatriation can limit the ability to scale up or down dynamically, potentially leading to overused or overburdened on-premises infrastructure. This lack of flexibility might result in inefficient resource allocation and increased costs.
The effect on cloud spending varies depending on factors such as an organization's size, workload characteristics and specific requirements. Before deciding on cloud repatriation, conduct a thorough cost analysis, and evaluate long-term spending and operational efficiency implications.
Cloud resource misconfiguration
Cloud misconfiguration is common in shadow IT projects or cloud initiatives that rely on the manual management of resources by inexperienced staff. A typical example is resources that run when a project no longer needs them, leading to unnecessary costs.
It's helpful for IT teams to understand how applications consume cloud resources. This process can be challenging for two reasons. First, new cloud services launch at a rate at which it's hard for any cloud engineer or solution architect to keep up. A lack of documentation, such as a detailed cloud services catalog, further contributes to this problem's prevalence. A second challenge is teams deploying additional technology to provide visibility into resource usage. This addition can result in overprovisioning resources to ensure performance and availability, while spending more on cloud than necessary.
Failure to optimize cloud resource usage contributes to cloud overservicing. Best practices, such as autoscaling, shutting down unused resources and implementing cost-efficient instance types, are foundational to managing cloud spending.
Lack of governance
When different teams and departments procure cloud services independently, the lack of coordination might result in overprovisioning. It can be natural to push for decentralization, especially in larger, more bureaucratic organizations where cloud adoption isn't always consistent across departments or projects. A lack of centralized governance and control framework can prove costly.
Shadow IT projects
Shadow IT projects create an issue closely related to lack of governance. For example, a department might pay for a shadow IT project through expense reimbursements. These projects contribute to an organization's inability to use the economies of scale that centralized cloud resource management ensures.
How to avoid cloud overservicing
Proactivity is critical to avoid cloud overservicing. Here are some strategies enterprises can put in place to prevent cloud overspending.
Monitor cloud resource utilization regularly
Organizations should regularly track their cloud resource usage to identify areas where they might be overspending. An easy place to start is with the resource monitoring tools cloud service providers offer, such as the following:
- Amazon CloudWatch.
- Microsoft Azure Monitor.
- Google Cloud Monitoring.
Teams should research monitoring tools, as some high-profile acquisitions in the cloud resource allocation market have occurred, such as IBM's acquisition of Turbonomic. Teams can use a cloud management platform (CMP) to monitor their cloud resource allocation. Piloting a new CMP enables cloud professionals to use its monitoring features and reporting to identify use cases and ensure the platform delivers what's necessary for resource monitoring. Cloud teams with existing CMPs should take the time to review the resource monitoring features to ensure their organizations get the most out of them.
Institute resource allocation policies
Organizations must establish resource allocation policies that align with their business goals. The following are some general resource allocation policy guidelines:
- Understand workload requirements. Analyze the workload characteristics and resource demands of the organization's applications or services running in the cloud.
- Define resource allocation objectives. This process includes maximizing resource utilization, ensuring fair sharing among users or departments, prioritizing critical workloads, and optimizing performance and responsiveness.
- Identify allocation strategies. Resource allocation strategies can include fixed allocations, dynamic allocations, priority-based allocations or resource-sharing models based on business objectives. Weigh the benefits and tradeoffs of each choice.
Instituting resource allocation policies means continuous monitoring and fine-tuning to ensure their effectiveness and alignment with an organization's current goals and objectives. Cloud teams should work with business stakeholders to make necessary adjustments for optimization.
Use cloud cost optimization tools
Several tools are available to help manage cloud spending, such as AWS Cost Explorer, Microsoft Cost Management and Google Cloud Billing. These tools can provide insights into usage patterns and help companies identify areas where they can save costs.
The third-party cloud optimization tool market is burgeoning. When selecting a cloud cost optimization tool, work from documented requirements, and take the time to conduct a pilot project to test the tool's analysis and reporting.
Consider reserved instances
Reserved instances from the major providers offer a discount on cloud resource usage when organizations commit to using a specific number of resources for a certain period. While this option requires some upfront work and analysis, it can be crucial to an organization's overall cloud cost-saving strategy.
Adopt serverless architecture
Serverless computing enables organizations to pay only for the resources used when executing code, which eliminates the need to provision resources beforehand. This can lead to significant cost savings, as companies pay only for what they use.