Active-standby TCP Synchronization for Hitless Switchover
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for hitless switchovers in active-standby systems, particularly involving TCP processes, are complex and error-prone due to difficulties in synchronizing incoming and outgoing data transmissions between active and standby applications, leading to challenges in accurately tracking the status of TCP connections and ensuring seamless transitions.
Innovation Solution
A method that involves an active TCP module receiving data packets, communicating them to an active application module, and then synchronizing this information with a standby application module through APP SYNC messages, allowing for accurate tracking and acknowledgment of data transmission states, thereby facilitating a hitless switchover.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the standby computer tracks all data from the active computer through continuous communication, then the switchover reliability is improved, but the system complexity and error-proneness increase due to difficulty in synchronizing incoming and outgoing TCP data
Solution Approach 1:
The patent segments the TCP data flow into distinct components: incoming TCP data, outgoing TCP data, and application data. By treating these as separate entities that can be independently tracked and synchronized, the system reduces the complexity of maintaining overall state synchronization during active-standby transitions.
Solution Approach 2:
The patent implements preliminary synchronization actions where the standby computer proactively receives and processes TCP data acknowledgments before actual data transmission occurs. This allows the standby to pre-establish the correct state, eliminating the need for complex real-time synchronization during switchover.
2Measurement precision
If the system implements comprehensive tracking of TCP transmission status, then the accuracy of switchover state transfer is improved, but the operational complexity increases
Solution Approach 1:
The patent implements feedback mechanisms where TCP acknowledgment messages are captured and used to update the standby computer's state. This feedback loop ensures that the standby accurately reflects the active computer's transmission status without requiring complex manual tracking or monitoring procedures.
Solution Approach 2:
The patent introduces intermediary components that mediate between the TCP protocol layer and the application layer, specifically designed to capture and transmit acknowledgment status. These intermediaries simplify the tracking process by handling the complexity of state monitoring automatically, making the system easier to operate.
3Loss of information
If the standby computer receives and processes all TCP data packets, then the data completeness is improved, but the processing time and potential for errors increase
Solution Approach 1:
The patent extracts only the essential synchronization information from TCP data packets—specifically acknowledgment status and sequence numbers—rather than processing complete data payloads. This extraction approach ensures data completeness for switchover purposes while significantly reducing processing time and minimizing error potential.
Data Source
AI summary
Embodiments of the invention include a method for maintaining an active-standby relationship between an active control card and a standby control card in a network element. The network element receives a data from a remote peer at the active control card. The network element communicates data from the active TCP module to an active application module in the active control card. The network element communicates synchronization data from the active application module to a standby application module on the standby control card. The network element communicates a single application synchronization acknowledgement from the standby application module to the active APP module. The network element communicates an application acknowledgment packet from the active application module to the active TCP module responsive to receiving the application synchronization acknowledgment. The network element then communicates an acknowledgement to the remote peer responsive to the application acknowledgement.