Cloud-Native Reconciliation for Network Equipment via CRD Translation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current cloud-native systems lack support for networking equipment, as they do not natively support desired state model reconciliation or detection of configuration drift in network entities, which are typically managed using YANG models and protocols like NETCONF.

Innovation Solution

The proposed solution involves generating cloud-native compatible resources from existing YANG models, converting these models to a format compatible with Kubernetes, and using asynchronous change notifications or periodic cycles to detect configuration drift and trigger reconciliation between the desired and observed states of network entities.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If YANG models are used to manage network entities, then network configuration management is achieved, but cloud-native integration and automated drift detection are not supported

Engineering Contradiction:
Improvecloud-native integrationVSAvoidautomated drift detection
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The patent introduces a translation layer that converts YANG models into cloud-native compatible formats (OpenAPI Schema/CRDs). This intermediary enables cloud-native systems to work with network entity configurations without requiring native YANG support, thereby achieving both cloud-native integration and maintaining the ability to detect configuration drift through standardized cloud-native mechanisms.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent transforms the data model parameters from YANG format to OpenAPI Schema format, changing the structural parameters while preserving the semantic meaning. This parameter transformation allows cloud-native orchestration systems to adopt and manage network configurations, enabling automated drift detection through standard cloud-native observability tools.

Inventive Principle:
Principle #35Parameter changes

2Productivity

If Kubernetes is used for cloud-native orchestration, then container management capability is improved, but network entity support is lacking

Engineering Contradiction:
Improvecontainer management capabilityVSAvoidnetwork entity support
Core Design Contradiction:
ProductivityVSAdaptability or versatility

Solution Approach 1:

The patent makes Kubernetes universally applicable to network entities by extending its resource management capabilities. Through the translation of YANG models to CRDs, Kubernetes can now manage both traditional containers and network entities (routers, switches, optical elements) through a unified interface, achieving multi-functionality without compromising its core container management strengths.

Inventive Principle:
Principle #6Universality (Multi-functionality)

3Loss of information

If XML encoding is used for YANG models, then data representation is achieved, but compatibility with cloud-native tooling is lost

Engineering Contradiction:
Improvedata representationVSAvoidcloud-native tooling compatibility
Core Design Contradiction:
Loss of informationVSAdaptability or versatility

Solution Approach 1:

The patent changes the encoding format parameter from XML to JSON-based OpenAPI Schema. This format transformation maintains complete data representation capability while achieving compatibility with cloud-native tooling, as JSON is the native format for Kubernetes and related cloud-native ecosystems, eliminating the verbosity and incompatibility issues of XML.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentUS12237975B2Cloud-native approach to support desired state model reconciliation with networking equipment
Publication Date: 2025.02.25 CIENA CORP
  • US12237975B2 patent drawing
  • US12237975B2 patent drawing
  • US12237975B2 patent drawing

AI summary

A method, implemented by a controller, includes steps of: subsequent to converting a bespoke model to Open Application Programming Interface (API) Schema that is Custom Resource Definition (CRD), receiving the CRD; receiving a target that is a data record that represents a network entity; receiving a configuration model instance for the target, wherein the configuration model instance includes one or more values that are compliant to the CRD and the one or more values represent a desired state of the network entity; receiving an observed state of the network entity; and determining drift between the observed state and the desired state.