Kubernetes Workload Proxying Across Incompatible Architectures
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing workload management systems struggle to manage workloads across computing architectures that are not compatible with Kubernetes, such as zSystems, limiting the use of Kubernetes' high availability and resilience features for non-compatible platforms.
Innovation Solution
A workload management system that deploys a proxy compute resource under a compatible architecture to interface with and control workloads on an incompatible architecture, ensuring a 1:1 relationship for seamless operation and resilience, including continuous monitoring and network traffic forwarding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a workload management system is designed for a specific computing architecture (e.g., Kubernetes for x86), then it can provide optimized management and high availability for workloads on that architecture, but it cannot manage workloads on incompatible architectures (e.g., zSystems)
Solution Approach 1:
The patent introduces a gateway component that acts as an intermediary between the Kubernetes control plane and workloads on incompatible architectures. The gateway translates Kubernetes API requests into architecture-specific commands, enabling cross-architecture workload management while maintaining the reliability of the original workload management system.
Solution Approach 2:
The system segments the workload management functionality into distinct components: the Kubernetes control plane for universal orchestration, architecture-specific adapters for translation, and workload agents running on the target architecture. This segmentation allows each component to specialize while working together to manage workloads across different architectures.
2Reliability
If workloads are distributed across multiple architectures to improve resource utilization, then availability increases, but the complexity of managing incompatible architectures increases
Solution Approach 1:
The gateway component is designed with universal functionality to handle multiple architecture types through a single interface. It can translate to and from various architecture-specific protocols while maintaining a consistent Kubernetes API interface, reducing the complexity of managing multiple incompatible systems.
Solution Approach 2:
The system implements health check mechanisms that continuously monitor workload status across different architectures and provide feedback to the Kubernetes control plane. This automated feedback loop simplifies management by enabling dynamic load balancing and failover without manual intervention.
Data Source
AI summary
A computer-implemented method of using a workload management system designed for workloads of a first computing architecture as well as for workloads of a second computing architecture is disclosed. The method comprises deploying a workload management control plane under a first computing architecture, identifying, by the workload management control plane, a requested remote workload only available for the second computing architecture, creating, by the workload management control plane, a proxy compute resource as interface for the requested remote workload, starting, by the proxy compute resource, the remote workload, controlling, by the proxy compute resource, an operation of the remote workload in a 1:1 relationship, by: upon receiving a termination signal, terminating the remote workload by the proxy compute resource and self-terminating the proxy compute resource, and upon a termination of the remote workload terminate the proxy compute resource.