Workload Deployment Optimization in Converged Infrastructure
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for deploying application workloads in converged infrastructure networks are inefficient due to a lack of visibility into the underlying physical network infrastructure, relying solely on virtualization layer metrics, which neglects compute components' performance and compliance in physical environments.
Innovation Solution
A method that calculates deployment optimization scores using metric data from individual converged infrastructure components, applying a weight-based algorithm to determine the ideal compute component for workload deployment, considering locality, health, capacity, and compliance metrics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If deployment decisions are based solely on virtualization layer metrics, then the deployment process is simplified and faster, but the visibility into physical infrastructure performance is lost leading to suboptimal deployment decisions
Solution Approach 1:
The patent extends the deployment optimization from the virtualization layer to the physical infrastructure layer by adding another dimension of monitoring. It collects metrics from physical components (network devices, storage systems, compute nodes) and integrates them with virtualization layer metrics, enabling deployment decisions that consider both abstracted and physical performance characteristics simultaneously.
Solution Approach 2:
The patent introduces a performance management system as an intermediary that bridges the virtualization layer and physical infrastructure. This intermediary collects, correlates, and analyzes metrics from both layers, providing comprehensive visibility without requiring direct exposure of physical infrastructure complexity to the deployment decision-making process.
2Reliability
If comprehensive physical infrastructure metrics are collected and analyzed, then deployment optimization is improved, but the system complexity and computational overhead increase
Solution Approach 1:
The patent segments the performance management system into multiple functional components: metric collection agents at physical devices, a central correlation engine, and deployment optimization modules. This segmentation allows distributed data collection while centralizing the complex analysis and decision-making logic, reducing the complexity burden on individual components.
Solution Approach 2:
The patent performs preliminary actions by pre-defining metric collection parameters, weightings, and optimization algorithms before deployment decisions are made. Performance thresholds and optimization criteria are established in advance, enabling automated decision-making without requiring complex real-time analysis during the actual deployment process.
3Ease of operation
If virtualization layer abstraction is used, then network management is simplified, but the underlying physical network infrastructure performance cannot be directly monitored
Solution Approach 1:
The patent introduces performance management agents as intermediaries that operate at the physical infrastructure level while reporting to the virtualization layer management system. These agents translate physical metrics into meaningful performance data that can be consumed by virtualization layer decision-making processes, maintaining abstraction benefits while enabling precise physical performance measurement.
Solution Approach 2:
The patent replaces direct physical monitoring mechanisms with a software-based metric collection and correlation system. Instead of requiring direct access to physical infrastructure components, the system uses standardized metric collection interfaces and data correlation algorithms to infer and measure physical performance characteristics from available data sources.
Data Source
AI summary
Methods, systems, and computer readable mediums for determining a system performance indicator representative of the overall operation of a network system are disclosed. According to one example, a method includes receiving an application workload for deployment into a network environment including a plurality of converged infrastructures and determining an overall deployment optimization score for each of the plurality of converged infrastructures. The method further includes determining a component optimization score for each of a plurality of compute components in a converged infrastructure belonging to the plurality of converged infrastructures that is associated with the highest overall deployment optimization score and deploying the application workload to a compute component belonging to the plurality of compute components that is associated with the highest component optimization score.