Workload Surge Handling via Dynamic Resource Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Software applications face challenges in handling workload surges, leading to unacceptably large response times or even becoming inoperable due to insufficient resource allocation during sudden increases in transaction volumes.
Innovation Solution
A system that correlates resource usage with workload signatures, uses machine learning to predict resource requirements, and triggers actions such as provisioning or throttling to manage capacity, ensuring the software application can handle unexpected workloads by dynamically allocating resources or deferring backend processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If static resource allocation is used for software application, then device complexity is reduced, but productivity deteriorates during workload surges
Solution Approach 1:
The patent implements dynamic resource allocation by continuously monitoring workload patterns and automatically adjusting resource provisioning based on detected surge conditions. The system transitions from static to dynamic resource management, enabling the software application to adapt its resource consumption in real-time according to actual demand patterns.
Solution Approach 2:
The system performs preliminary analysis of workload patterns and triggers proactive resource allocation before actual surges occur. By predicting potential workload increases based on historical data and patterns, the system pre-provisions resources to prevent performance degradation rather than reacting after the surge begins.
2Productivity
If resource allocation is increased to handle workload surges, then productivity is improved, but use of energy increases
Solution Approach 1:
The system dynamically adjusts resource allocation parameters based on detected workload conditions. By monitoring metrics such as transaction volume, response time, and pattern recognition, the system modifies resource provisioning parameters to match actual demand, avoiding both over-provisioning and under-provisioning scenarios.
Solution Approach 2:
The patent implements feedback mechanisms that continuously monitor system performance and workload patterns. This feedback loop enables the system to learn from actual usage patterns and refine its resource allocation decisions, ensuring energy-efficient provisioning that matches real demand rather than relying on conservative static allocations.
Data Source
AI summary
According to an aspect of the present disclosure, a correlation data correlating resource usage with workload signatures is maintained, each workload signature representing a cluster of block signatures, each block signature characterizing the transaction instances initiated in a corresponding block duration. For the transactions received in a current block duration, if a current block signature is not contained in the correlation data and if the current transaction arrival rate (TAR) is greater than an expected TAR, a resource requirement for the current block signature is computed. Actions to manage capacity to handle transaction instances are triggered if the resource requirement is greater than the resource allocation in the current block duration. As an unknown current block signature and a higher TAR may be indicative of a workload surge, triggering suitable actions for such block signatures facilitates such surges to be handled by the software application.