Control Plane Configuration for Edge Computing Footprint Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional computing systems face resource underutilization and space constraints in edge deployments due to the large footprint of control planes in container orchestration architectures, which limits the number of worker nodes that can be accommodated.
Innovation Solution
A container orchestration platform utilizing a pair of control switches, operating in active and passive modes, to manage worker nodes and workloads, with a virtual IP address for communication and a key-value database for storing data, enabling fault tolerance, high availability, and scalability while minimizing the physical footprint.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a traditional control plane architecture is used in container orchestration, then fault tolerance and high availability are achieved, but the physical footprint becomes large, limiting the number of worker nodes that can be accommodated
Solution Approach 1:
The control plane is segmented into multiple control switches instead of using a single monolithic control plane. Each control switch handles a subset of worker nodes, distributing the control plane's functionality across multiple smaller units. This segmentation reduces the physical footprint of each individual control component while maintaining the overall fault tolerance through redundancy and distribution.
Solution Approach 2:
The patent transitions from a single-layer control plane architecture to a multi-dimensional hierarchy where control switches are organized in layers. Control switches can operate at different levels (e.g., edge control switches, core control switches), creating a dimensional structure that reduces the footprint at any single level while preserving fault tolerance through the hierarchical redundancy.
2Productivity
If more worker nodes are accommodated in edge deployments, then resource utilization improves, but space constraints are exceeded due to the large control plane footprint
Solution Approach 1:
By segmenting the control plane into multiple smaller control switches, the space required for control functionality is distributed and reduced at any single location. This enables more worker nodes to be accommodated within the same physical space constraints, thereby improving resource utilization without exceeding available space.
Solution Approach 2:
The patent uses multiple control switches that can be deployed in replicated configurations across different locations or racks. Instead of concentrating all control plane functions in one large unit, replicated smaller control switch units are distributed, reducing the footprint per worker node while maintaining adequate control coverage for high resource utilization.
3Device complexity
If control plane functions are concentrated in a single location, then management is simplified, but the system becomes vulnerable to single points of failure
Solution Approach 1:
The control plane is divided into multiple control switches that can be managed independently. While this increases the number of components, each control switch has a well-defined scope and can be managed separately, reducing the complexity of managing any single control component. The segmentation also eliminates single points of failure, as the failure of one control switch does not bring down the entire control plane.
Solution Approach 2:
Each control switch is designed with local quality principles, where each unit has specialized functions and can operate semi-independently. This allows for localized management and troubleshooting of individual control switches while maintaining overall system coherence. The local quality approach enables simplified management of each component while achieving system-wide fault tolerance through the distributed architecture.
Data Source
AI summary
Examples relate to a system comprising a cluster of worker nodes communicatively connected to a control plane. The control plane includes a first control switch and a second control switch. The control switches are configured with a virtual IP address. The IP address is associated with a control switch operating in an active mode. The control switches communicate with a cluster of worker nodes via the virtual IP address. A single instance of a key-value database is deployed on one of the first control switch and the second control switch that is operating in active mode, and the first control switch and the second control switch are functionally coupled by a distributed block system for data synchronization.