Cloud Infrastructure Benchmarking via Executable Code
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cloud infrastructure monitoring systems lack efficient methods for performing generalized tests of performance and capacity issues related to virtual machines and containers from a centralized location, which can lead to violations of Service Level Agreements (SLAs) due to inadequate assessment of resource utilization patterns.
Innovation Solution
A method involving the use of benchmark messages that contain program code or references to code, executed on monitored entities within the cloud infrastructure to assess capabilities, with the results providing execution time and optional results, enabling a more accurate evaluation of virtual machine or container performance and resource utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional monitoring systems are used to track resource utilization, then basic resource tracking is available, but accurate assessment of performance and capacity issues cannot be performed, leading to SLA violations
Solution Approach 1:
The system performs preliminary benchmark testing by executing standardized programs on monitored entities before actual workloads are deployed. This preliminary action establishes baseline performance characteristics and capacity limits, enabling accurate assessment of performance and capacity issues that prevent SLA violations during subsequent operations.
Solution Approach 2:
The system continuously monitors resource utilization patterns and feeds this information back to establish more accurate performance models. By analyzing actual workload behavior against benchmark results, the system refines its assessment capabilities, improving both measurement precision and SLA compliance through iterative feedback loops.
2Ease of operation
If centralized monitoring is implemented to test performance from a single location, then coordination is simplified, but the complexity of executing and managing benchmark programs increases
Solution Approach 1:
Monitored entities automatically execute benchmark programs and return performance data without requiring manual intervention or complex configuration from the centralized monitoring system. The entities self-manage their own benchmarking process, reducing the operational burden on centralized systems while maintaining coordination benefits.
Solution Approach 2:
The monitoring system employs universal benchmark programs that can be executed across diverse monitored entities (virtual machines, containers, physical servers) with different configurations. This multi-functionality allows centralized coordination while simplifying management, as the same benchmark framework adapts to various entity types without requiring entity-specific customization.
Data Source
AI summary
A compute infrastructure interconnected with a network configured for allotting a benchmark message. The benchmark message comprises program code, or references to program code, to be executed on a reflecting entity. The reflecting entity is able to execute program code. The program code requests capabilities of the reflecting entity when being executed. The compute infrastructure is further configured for sending of the benchmark message to the reflecting entity. The compute infrastructure is further configured for receiving the benchmark message and for executing, in the reflecting entity, the program code. The compute infrastructure is further configured for providing a benchmark reply message. The benchmark reply message comprises execution time and optionally execution result of the execution of said program code. The compute infrastructure is further configured for returning the benchmark reply message from the reflecting entity.