Hitless Switch Management Module Failover via State Synchronization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional failover methods between primary and backup switch management modules in switched network elements are not hitless, resulting in packet loss and inefficiencies during software upgrades due to the need for complete initialization and protocol restarts, which can lead to dropped packets and compatibility issues between different software versions.
Innovation Solution
Implementing a master-slave mode communication system where the backup switch management module continuously monitors the primary module, receives and stores protocol state and forwarding data structures, and can seamlessly take over as the new primary without resetting packet forwarding hardware, using a canonical message format and bracketing for atomic transactions to ensure synchronization and compatibility across software versions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the backup switch management module performs complete initialization and protocol restart from initial state during failover, then the new primary module can participate in routing and forwarding protocols, but packets are dropped and network devices may mark the failed device as unavailable due to protocol timeout values
Solution Approach 1:
The backup management module continuously monitors the primary module's operation and pre-loads configuration information, routing and forwarding protocol states, and forwarding tables into its memory before failover occurs. This preliminary action enables the backup module to immediately assume the primary role without performing complete re-initialization, thereby reducing failover time and preventing packet loss while maintaining protocol compliance.
2Reliability
If the backup switch management module restarts network protocols from initial state during failover, then the new primary module can establish proper protocol states and data structures, but packets are dropped during the time required to participate in protocols and build data structures
Solution Approach 1:
The backup management module creates a copy of the primary module's operational state, including routing and forwarding protocol states and forwarding tables, and loads this copy into its own memory. This copying approach allows the backup module to assume the exact protocol state of the primary module without re-establishing protocols from scratch, thereby avoiding packet drops while maintaining protocol state consistency.
3Adaptability or versatility
If the backup switch management module uses a new software version during failover, then software upgrades can be performed, but data structures between different software versions may not be compatible causing the new version to be unable to use data from the prior version
Solution Approach 1:
The system employs a canonical message format as an intermediary layer for communicating data structures between management modules running different software versions. This canonical format acts as a mediator that translates and standardizes data structure representations, enabling compatible data exchange between modules with different software versions while maintaining data integrity and structural consistency.
4Adaptability or versatility
If the backup switch management module initializes hardware and participates in forwarding protocols to build forwarding databases during software upgrade, then the new software version can operate independently, but the process is not hitless due to the time required to build appropriate data structures
Solution Approach 1:
During software upgrade, the backup management module pre-loads the new software version and preserves the operational state data structures from the primary module in its memory before failover. This preliminary preparation allows the new software version to immediately assume the primary role with existing data structures already in place, eliminating the need to rebuild forwarding databases from scratch and achieving hitless upgrade.
Data Source
AI summary
Methods and systems for hitless switch management module failover and upgrade are disclosed. According to one method, a master switch management module participates in network protocols and performs packet forwarding operations. The master switch management module distributes protocol state and packet forwarding information to the slave switch management module. The slave switch management module continuously monitors the operational state of the master switch management module. In response to detecting failure of the master switch management module or a forced failover initiated by the user interface on the master switch management module, the slave switch management module begins network protocol operation in the master mode in a state where the master switch management module last operated correctly.