MEC Application Testing With Clustered Parameter Tuning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing software application performance in multi-access edge computing environments is challenging due to variations in infrastructure and network conditions, leading to inefficiencies and suboptimal user experiences.
Innovation Solution
An orchestrator system that analyzes and adjusts software application performance parameters by comparing similar edge computing environments, using machine learning and predictive models to optimize performance across geographically dispersed networks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If software application parameter values are adjusted to improve performance on target computing environments, then performance is improved, but testing and optimization time increases
Solution Approach 1:
The system performs preliminary testing and parameter optimization on test computing environments before deploying changes to target environments. By conducting AB testing and clustering analysis in advance on representative test environments, the system prepares optimal parameter configurations beforehand, reducing the time needed for live environment adjustments and iterations.
Solution Approach 2:
The system creates virtual copies of target computing environments as test environments that replicate infrastructure and network conditions. By testing parameter adjustments on these copied environments first, the system can evaluate performance impacts without affecting production systems, enabling safe optimization experiments that reduce overall optimization time.
2Reliability
If comprehensive testing is performed across multiple computing environments to ensure reliability, then reliability is improved, but system complexity increases
Solution Approach 1:
The system automatically selects appropriate test computing environments and determines optimal parameter adjustments without requiring manual configuration of complex testing frameworks. The environment selection and parameter optimization processes are self-managing, reducing the complexity burden on operators while maintaining comprehensive testing across multiple environments.
Solution Approach 2:
The system focuses testing efforts on specific parameter variations that have the most significant impact on performance and reliability. By identifying and testing only the critical parameter changes rather than exhaustively testing all possible configurations, the system achieves reliable optimization results with reduced testing system complexity.
3Productivity
If parameter adjustments are made based on test environment results, then performance is improved, but adaptability to different computing environments may be reduced
Solution Approach 1:
The system tailors parameter adjustments to specific target computing environments by using clustering analysis to group environments with similar characteristics. Each environment receives customized parameter optimizations based on its unique infrastructure and network conditions rather than applying uniform changes, thereby maintaining high adaptability while achieving performance improvements.
Solution Approach 2:
The system continuously monitors performance metrics from target computing environments and uses this feedback to refine future parameter adjustments. By incorporating real-world performance data into the optimization process, the system adapts parameter settings to actual environmental conditions, maintaining versatility across different deployments while achieving sustained performance gains.
Data Source
AI summary
Methods, computer program products, and systems are presented. The method computer program products, and systems can include, for instance: testing a software application having instances that run on (a) a test computing environment and (b) a target computing environment, wherein each of the test computing environment and the target computing environment is provided by a mobile access edge computing (MEC) environment, wherein the testing includes varying a software application performance impacting parameter value impacting performance of an instance of the software application running on the test computing environment, and examining metrics data resulting from the varying the software application performance impacting parameter value; providing, in dependence on the metrics data, an action decision to adjust a software application performance impacting parameter value impacting performance of a software application instance of the software application running on the target computing environment; and adjusting the software application performance impacting parameter value impacting performance of the software application instance of the software application running on the target computing in accordance with the action decision.