Usage Management and Optimization

Existing workload usage prediction and management

Periodic usage increase prediction and management

Generally, as cloud usage increases, costs also increase.

Except for special economic conditions, most businesses grow over time and data continuously accumulates, leading to increased demand for cloud resources.

If these resource demands are not managed properly, costs are likely to become uncontrolled and increase.

Therefore, it is necessary to periodically forecast cloud usage and take appropriate actions accordingly.

Definition of Review Cycle

First, define the review cycle for cloud resources, and allocate time and resources according to that cycle to achieve cost optimization.

When setting the review cycle, you can refer to the discount commitment period of cloud resources.

For example, Samsung Cloud Platform offers contract discounts, Cost Savings, and Planned Compute in 1‑year or 3‑year terms.

By setting a resource adjustment cycle of three months or six months based on the expiration date of these agreements, it becomes easier to develop resource operation plans.

Additionally, aligning the review cycle with the company’s fiscal period allows it to be effectively used for performance evaluations by department.

When setting the review cycle, you should also consider the software license agreement, the operations management contract term, and similar factors.

For example, you can set a review schedule where the entire resources are reviewed every 12 months and the data store every 6 months.

Usage Increase Prediction

The factor that has the greatest impact on cloud costs is increased usage.

We need to adjust resources to reflect the impact of increased usage on the currently operating cloud environment.

You can perform prediction through the following procedure.

1. Evaluate whether all expense items are properly classified according to the organization’s departments and processes, and adjust them as needed.

2. Calculate the average expenditure for each group over the past three months.

3. Calculate the average expenditure compared to the same period one year ago.

4. Compare the two averages to identify the trend and reflect each group's growth rate.

5. Adjust resource quantities and pricing plans to reflect future plans.

6. Share with the FinOps stakeholder and coordinate feedback.

In the fifth step’s plan integration process, adjust the pricing plan and modify the quantity to reflect any newly increased or decreased resource usage.

The related FinOps cost modeling is covered in detail in III. FinOps Strategy Development and Execution 1.3 Cost Modeling.

Review of new services, features, and configurations

Most cloud providers, including Samsung Cloud Platform, are continuously adding new technologies and services.

Some of these enable new business experiments, while others contribute to improving the performance of existing resources.

To keep the workload cost-effective, you should regularly review the feasibility of adopting new services, features, and components.

Forecasting and Managing Usage Based on External Influences

Analyzing external factors that affect cloud usage is also important.

To do this, we understand the patterns and characteristics of computing tasks, define response time as a key performance indicator, and assess whether demand fluctuates.

Additionally, analysis of the predictability, repeatability, rate of change, and scale of external influences is also required.

Set the analysis period based on a sufficient duration (at least one year) to account for seasonality (Seasonality).

Through this analysis, you can adjust resources based on the predicted impact and evaluate the resulting cost-effectiveness.

• Workload Type Identify the business type of the system under analysis. For example, e-commerce, internal business systems, and machine learning services have different performance requirements and resource characteristics, so we identify the required resources through workload analysis.

Usage Rate and Performance Metrics Analyze changes in resource usage and develop resource adjustment strategies for maximum and minimum utilization. We also use performance metrics such as response time and latency at this point.

Request Load Type Analyze the traffic pattern of the request load and determine the direction of resource adjustment based on whether the workload is centered on database transactions or on content delivery.

You can perform the following tasks to make decisions required for usage forecasting and resource adjustment.

• We gain insights into workloads by using log files and monitoring data extracted from monitoring tools such as Cloud Monitoring. Obtain data on periodic changes and assess trends in demand variability and growth patterns.

• Collaborate with departments that can affect demand to verify whether an event occurs.

Resource Optimization

Resource type, size, quantity adjustment

Resource optimization is essential for balancing the two goals of cost reduction and service reliability.

Resizing resources to optimize size and quantity requires strategic, data-driven decision-making rather than simple technical adjustments, and the items to review are as follows.

Resource attributes and cost considerations By optimally selecting the type, size, and quantity of resources, you can meet technical requirements at minimum cost. When performing resizing for cost optimization, you must comprehensively consider all resources included in the workload, each resource’s attributes, and even the labor costs required for the adjustment. If the labor cost required for resizing exceeds the amount that can be saved, it is advisable to perform the task as a one‑time operation at the point of service change or discontinuation rather than repeating it regularly. To adjust resource sizes, you must obtain visibility into which resources are currently being used and to what extent. This visibility includes CPU usage, memory usage, network throughput, and disk usage, and based on this data you can define the server type and disk capacity.

• Resizing for resource optimization Resizing should not be performed solely for cost reduction; it must be done carefully to avoid negatively impacting service operations. The operations team’s primary goal is to reliably maintain the operational capacity required for the service. In particular, adjusting the size of resources that support commercial applications is highly complex due to licensing issues and therefore requires a careful approach. Resources can be low-cost or high-cost, and focusing optimization efforts on high-cost resources can be more efficient. As mentioned earlier, if the labor cost involved in resizing exceeds the resource cost that can be saved, the task may be inefficient. It is advisable to decide on resizing by considering the limits of the amount that can be saved, and focusing on higher-spec databases rather than lower-spec Virtual Servers can yield greater cost savings.

Data-driven Cost Optimization Strategy Data-driven resizing can be used for capacity adjustments in a Scale-up approach, and a Scale-out approach can also be considered. This approach is a concept for scaling computing Node, allowing both manual and automatic scaling strategies to be applied. Examples of manual adjustments include increasing the number of Virtual Servers running on a Load Balancer, or adjusting the number of Node pools or pod replica sets in a Kubernetes Cluster. Automation is also possible by using metrics such as average CPU utilization for these tasks. By specifying the minimum and maximum number of worker nodes, you can set the capacity range of computing resources required to handle the workload, and achieve cost optimization by dynamically adjusting capacity based on metrics.

Idle resource disposal

Among the tasks required for resource optimization in the cloud, the most important and effective task is idle resource management.

Resources can be provisioned out of necessity, yet after some time they may become idle resources that are no longer used and remain without being deleted or scaled down.

These measures for idle resources are essential for reducing cloud costs and represent the most effective solution.

Idle Resource Management Procedure In cloud environments, unused idle resources can arise over time. These resources should be deleted, but if they are not managed, they can be neglected due to administrator negligence or lack of visibility into resource management. If the resource was not created directly by the administrator, deletion may be difficult, and it may contain important data or be retained for a specific purpose. Therefore, to discard idle resources, a method to verify the usage history of the resources and a disposal procedure are required.

Idle Resource Lifecycle Management First, you need to manage tags related to the resource lifecycle. From a lifecycle perspective, the resource must identify whether it is intended for testing and when testing is completed. If the resource uses a license with an expiration period, the administrator must also be able to identify that information. This information can be implemented through tags, allowing you to establish tag policies related to the resource lifecycle to define the resource’s purpose, expiration period, etc. After that, a process for resource disposal must be established, and all resources must include information about the relevant department and responsible personnel. Before disposing of resources, you must notify stakeholders in advance and prevent data loss through a verification process. If resources are managed according to task priority, it is also possible to implement automatic disposal or deletion automation for low‑priority resources.