Workload Placement Using Network Drift Detection in Scalable Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The performance of computer-implemented services in data processing systems is impaired by limitations in network connectivity and resource availability, which hinder the optimal operation of hardware components.
Innovation Solution
A workload manager selects locations for workload performance based on resource availability and network connectivity, using optimization testing to identify the most conducive network characteristics and computing resources, thereby facilitating optimal workload execution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If workload is assigned without considering network connectivity and resource availability, then device complexity is reduced, but service performance deteriorates
Solution Approach 1:
The system performs preliminary optimization testing to establish associations between workload types and network characteristics before actual workload assignment. This pre-computed knowledge base enables subsequent workload assignments to be made efficiently without real-time complex analysis, thus improving service performance while maintaining manageable complexity.
Solution Approach 2:
The system implements feedback mechanisms by continuously monitoring network connectivity and resource availability, and using this information to dynamically adjust workload assignments. The optimization testing results feed into the workload manager's decision-making process, creating a closed-loop system that adapts to changing conditions and maintains optimal performance.
2Reliability
If hardware components operate without optimal network connectivity, then network connectivity requirements are simplified, but hardware component operation deteriorates
Solution Approach 1:
The system applies local quality by matching specific workload types with their optimal network characteristics. Different workload types have different network requirements, and the system ensures each workload is assigned to locations with appropriate network connectivity qualities, thereby ensuring reliable hardware operation without requiring all components to adapt to all network conditions.
Solution Approach 2:
The system changes network parameters (connectivity quality, bandwidth, latency) as selection criteria for workload assignment. By varying these parameters based on workload type and optimizing the match between workload and network conditions, the system ensures hardware components operate reliably under their specific network requirements rather than forcing uniform adaptability across all components.
3Measurement precision
If optimization testing is performed for all workload types, then association accuracy improves, but testing time increases
Solution Approach 1:
The system performs optimization testing selectively rather than exhaustively for all possible workload types. It conducts testing for representative workload types or updates associations periodically and incrementally, achieving sufficient association accuracy without the prohibitive time cost of comprehensive testing of every workload type in every condition.
Data Source
AI summary
Methods, systems, and devices for managing operation of a data processing system that provides computer implemented services are disclosed. To do so, a request to perform a workload may be obtained. A type of the workload may be identified based on the request, and based on this type, computing resources required for performance of the workload may be identified. Potential locations for the performance of the workload may be filtered based on the required computing resources to identify a portion of the potential locations. Based on associations between types of workloads and sets of network characteristics, a set of network characteristics may be identified for the type. Based on the set of network characteristics, the potential locations may be further filtered to identify a sub-portion of the potential locations. Based on this identification, performance of the workload by at least one data processing system of the sub-portion may be initiated.