Multi-Cluster Edge System with Intelligent Scheduler
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing edge computing systems face challenges in managing multiple clusters, optimizing resource allocation, and enabling high-speed collaboration between cloud, edge, and near-edge terminals, with limitations in vertical and horizontal collaboration, container performance, and network architecture.
Innovation Solution
A multi-cloud edge system is proposed, featuring a core cloud, multi-cluster edge nodes, and near-edge nodes connected via a tunneling protocol, with a shared storage device and intelligent scheduler for optimized resource management and collaboration, supporting monolithic applications, micro-services, and FaaS, and utilizing in-memory container storage for high-performance data processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a single cluster Kubernetes environment is used, then the system is simple to manage, but it cannot support distributed computing across multiple clusters
Solution Approach 1:
The system divides the computing environment into multiple independent Kubernetes clusters, each managing its own resources and workloads. This segmentation allows each cluster to operate autonomously while contributing to the overall distributed computing capability, resolving the contradiction between multi-cluster support and system simplicity.
Solution Approach 2:
A service mesh (Istio) is introduced as an intermediary layer between multiple Kubernetes clusters. This mediator enables communication and coordination between clusters without requiring direct complex configurations between them, thus providing multi-cluster support while maintaining manageable system complexity through standardized interfaces.
2Speed
If data is processed at centralized cloud data centers, then resource management is simplified, but transmission delay increases for edge terminals
Solution Approach 1:
The centralized cloud processing architecture is segmented into distributed edge computing nodes deployed closer to terminals. This segmentation enables data processing to occur at multiple locations (cloud, edge, near-edge) rather than a single centralized location, reducing transmission delay while managing complexity through a hierarchical network architecture.
Solution Approach 2:
The system adds a spatial dimension to the computing architecture by deploying edge and near-edge nodes at different physical locations closer to terminals. This dimensional expansion from centralized to distributed architecture reduces transmission distance and delay, while the structured multi-layer design manages the resulting network complexity.
3Speed
If traditional network connections are used between clusters, then implementation is straightforward, but connection speed and performance are insufficient
Solution Approach 1:
A service mesh acts as an intermediary network layer that establishes optimized communication paths between clusters. This mediator provides high-speed connections through standardized protocols and routing mechanisms, achieving fast inter-cluster communication without requiring complex direct network configurations between each pair of clusters.
4Productivity
If resources are allocated dynamically without optimization, then resource flexibility is high, but resource utilization efficiency decreases
Solution Approach 1:
An intelligent scheduler is implemented that continuously monitors resource usage, workload demands, and system state across clusters. This feedback mechanism enables dynamic optimization of resource allocation by adjusting scheduling decisions based on real-time conditions, improving resource utilization efficiency while managing complexity through automated control loops.
Solution Approach 2:
The system performs preliminary resource provisioning and pre-allocation based on predicted workload patterns and historical data. This preliminary action prepares resources in advance, reducing the need for complex real-time scheduling decisions and improving overall resource utilization efficiency through proactive rather than reactive allocation.
Data Source
AI summary
Disclosed herein is a multi-cloud edge system. The multi-cloud edge system includes a core cloud, a multi-cluster-based first edge node system, and a multi-cluster-based near edge node system, wherein the multi-cluster-based first edge node system includes multiple worker nodes, and a master node including a scheduler.