Kubernetes Node Customization for Network Function Deployment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for migrating and deploying network functions across different servers or virtual machines are time-consuming and lack a consistent, vendor-agnostic approach to capture software requirements, leading to inefficiencies and downtime.
Innovation Solution
The Telco Cloud Automation (TCA) system facilitates customization of network function infrastructure by parsing descriptors to configure Kubernetes nodes dynamically, enabling on-demand resource allocation and late binding of hardware and software features based on specific requirements, without swapping underlying servers or virtual machines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If network functions are migrated and deployed across different servers or virtual machines using existing methods, then network function deployment is achieved, but the process is time-consuming and causes downtime
Solution Approach 1:
The system performs preliminary configuration of Kubernetes nodes by parsing function descriptors and pre-configuring infrastructure requirements before network function deployment. This includes pre-configuring hardware features, software features, and resource allocations based on the function descriptor, so that when deployment occurs, the node is already ready to accept the network function without time-consuming configuration steps
Solution Approach 2:
The TCA system enables automated self-service deployment by automatically parsing function descriptors, identifying infrastructure requirements, selecting appropriate Kubernetes nodes, and configuring them without manual intervention. The system autonomously handles the entire deployment process including resource allocation, node selection, and configuration application, eliminating the need for manual server swapping or virtual machine migration operations that cause downtime
2Extent of automation
If manual configuration methods are used for network function deployment, then deployment can be performed, but the process lacks automation and consistency
Solution Approach 1:
The system replaces manual mechanical configuration processes with automated software-based configuration. Instead of manually configuring servers or virtual machines, the TCA system uses automated parsing of function descriptors and programmatic configuration of Kubernetes nodes through APIs and configuration management tools, substituting human-operated mechanical processes with automated software systems
Solution Approach 2:
The system dynamically changes configuration parameters based on the parsed function descriptor. It automatically adjusts hardware features, software features, resource allocations, and node selection criteria according to the specific requirements defined in the function descriptor, enabling consistent automated deployment across different network functions with varying parameters
3Adaptability or versatility
If existing deployment methods are used, then network functions can be deployed, but they lack a consistent vendor-agnostic approach to capture software requirements
Solution Approach 1:
The system implements a universal vendor-agnostic approach by designing the TCA platform to parse and process function descriptors from multiple vendors through a common interface. The system can handle different vendor-specific formats and requirements by translating them into a standardized internal representation, allowing the same deployment mechanism to work across different vendors without vendor-specific customization
Solution Approach 2:
The system introduces an intermediary layer in the form of the function descriptor parsing mechanism that sits between vendor-specific deployment requirements and the underlying Kubernetes infrastructure. This intermediary translates various vendor formats into a standardized configuration format, ensuring consistent and accurate deployment across different vendors while maintaining precision in capturing software requirements
Data Source
AI summary
Some embodiments provide a method that generates (i) a node profile of a worker node in a workload cluster for deploying a first network function and (ii) a host profile of a virtual machine that implements the workload cluster and a management cluster that manages the workload cluster. The method updates a configuration map of the worker node based on the node profile. The method uses a node configuration operator in a remote data center to configure the worker node based on the configuration map. The method uses a virtual machine configuration operator in the remote data center to configure one or more nodes of the management cluster based on the host profile.