Weight-Based EMS Failover in Container Multi-Cluster Environments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Container-based multi-cluster environments face challenges in maintaining efficient redundancy operations and managing state-dependent applications due to issues like data synchronization delays, data collisions, and the lack of support for active-standby states, leading to potential network element dispersion errors and service disruptions.
Innovation Solution
A method and apparatus for controlling container applications in a multi-cluster environment by using a service weight-based approach, where a first EMS identifies an abnormal state and transmits a state change message to a second EMS to switch from standby to active, enabling real-time synchronization and immediate application execution in a secondary cluster.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data synchronization is implemented between clusters, then service continuity is improved, but synchronization delays and data collisions occur
Solution Approach 1:
The system performs preliminary actions by pre-synchronizing data between clusters before failure occurs. The active-standby cluster configuration maintains pre-synchronized data ready for immediate takeover, eliminating the need for real-time synchronization during failure recovery, thus avoiding synchronization delays.
Solution Approach 2:
The patent introduces an intermediary mechanism (the standby cluster with pre-synchronized data) that mediates between the active cluster and service continuity requirements. This intermediary allows seamless failover without direct real-time synchronization during the failure event, preventing data collisions while maintaining service continuity.
2Productivity
If active-standby state support is implemented, then redundancy efficiency is improved, but state management complexity increases
Solution Approach 1:
The system segments the cluster environment into distinct active and standby states with clearly defined roles and responsibilities. This segmentation simplifies state management by creating discrete, manageable states rather than continuous complex state transitions, improving redundancy efficiency while controlling complexity.
Solution Approach 2:
The patent implements dynamic state transition capabilities that allow the system to adapt between active and standby states based on operational conditions. This dynamic approach enables efficient redundancy management through automated state changes, reducing manual intervention complexity while maintaining high productivity.
3Reliability
If real-time synchronization is implemented, then service disruption is reduced, but network element dispersion errors increase
Solution Approach 1:
The system performs data synchronization in advance during normal operation, so that when failure occurs, the standby cluster already has the necessary data ready. This preliminary action eliminates the need for real-time synchronization during failure events, thereby preventing network element dispersion errors while still achieving minimal service disruption.
Data Source
AI summary
The present disclosure relates to a container-based cluster environment and, to a method and apparatus for controlling a container application in a container-based multi-cluster environment. The method performed by a first element management system (EMS) in a first cluster comprises: if the first EMS is in an active state, identifying pods that are running and in service; identifying a weight for each of the pods that are running and in service; based on the weight for each of the pods that are running and in service, identifying that a state of the first EMS is an abnormal state requiring state transition; if the state of the first EMS is the abnormal state, transmitting, to a second EMS in a second cluster, a state change message instructing to change a state of the second EMS from a standby state to the active state; and changing the state of the first EMS to the standby state, wherein the active state includes a state in which one or more service pods included in an EMS are running and in service, and the standby state may include a state in which the one or more service pods are running but are not in service.