TELCOCOST02-BP02 Choose the most efficient compute resource for your Network Function
When implementing cloud-based network functions (CNFs), it is important to benchmark your applications on different compute types to determine the optimal balance of performance and cost-efficiency. CNFs have diverse compute requirements, so gravitating to the lowest cost option may result in poor performance. Older compute architectures and hardware can hamper efficiency and drive-up expenses over time.
Desired outcome:
-
Identify the most cost-effective compute resources that meet the performance requirements of your cloud-based network functions (CNFs).
-
Optimize the balance between cost and performance for your CNF workloads.
-
Avoid over-provisioning or under-provisioning of compute resources, which can lead to increased costs.
Common anti-patterns:
-
Selecting compute resources solely based on the lowest cost, without considering performance requirements.
-
Relying on outdated compute architectures and hardware that are less efficient and cost-effective.
-
Lack of benchmarking and profiling of CNF workloads on different compute options.
Benefits of establishing this best practice:
-
Significant cost savings by selecting the most optimal compute resources for your CNFs.
-
Improved CNF performance and reliability by matching compute capabilities to workload needs.
-
Enhanced operational efficiency by right-sizing compute resources to avoid over-provisioning.
-
Increased agility in responding to changing compute requirements for your CNF workloads.
Level of risk exposed if this best practice is not established: High
Implementation guidance
When implementing cloud-based network functions (CNFs), it is crucial to carefully evaluate the compute resource options to strike the right balance between cost and performance. CNFs often have diverse compute requirements, and simply selecting the lowest-cost compute option may result in poor performance and sub-optimal outcomes.
By benchmarking your CNF applications on different compute types, you can identify the most efficient resources that meet your performance needs. This may involve evaluating various AWS compute services, such as Amazon EC2 instances with different processor architectures, memory configurations, and storage options. Older compute architectures and hardware can also hamper efficiency and drive-up expenses over time, so it is important to consider the long-term costs and benefits of your compute choices.
The goal is to optimize the cost-performance tradeoff for your CNF workloads, verifying you are not over-provisioning or under-provisioning compute resources, which can lead to increased costs.
Implementation steps
-
Profile the performance and resource utilization characteristics of your cloud-based network functions (CNFs).
-
Benchmark your CNF applications on a variety of AWS compute instances, including different processor architectures, memory configurations, and storage options.
-
Analyze the performance and cost data to identify the most efficient compute resources that meet your CNF's requirements.
-
Consider the long-term implications of your compute choices, evaluating factors like energy efficiency, hardware lifecycle, and future performance trends.
-
Implement mechanisms to dynamically adjust the compute resources allocated to your CNFs based on changing demands and workload characteristics.
-
Continuously monitor and optimize your CNF compute resource allocations to verify cost-effectiveness while maintaining performance.
Resources
Key AWS services:
