Gateway Switchover Mechanism for Data Center Traffic Continuity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In data centers, planned maintenance of virtual network gateways can cause interruptions and data loss due to delays in switchover processes, especially when the standby gateway instance does not assume responsibility before the primary instance goes offline, leading to lost revenue and customer dissatisfaction.
Innovation Solution
Implementing a forwarding mechanism where the active gateway instance delays notification of its offline status to allow sufficient time for the backup instance to establish connections, and queuing data packets to prevent loss during the switchover, minimizing tunnel reconnect times and packet loss.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the standby gateway instance waits for health probe confirmation before taking over, then traffic forwarding reliability is improved, but switchover time increases causing service interruption
Solution Approach 1:
The standby gateway instance performs preliminary actions by pre-establishing tunnel connections to backend virtual machines before the switchover is triggered. When maintenance is initiated, the standby instance already has active tunnels ready, allowing immediate traffic forwarding without waiting for health probe confirmations or tunnel reconnection processes.
2Productivity
If the primary gateway instance goes offline immediately for maintenance, then maintenance efficiency is improved, but data packet loss occurs during switchover
Solution Approach 1:
The system performs preliminary actions by having the standby gateway instance pre-establish tunnel connections and prepare for traffic forwarding before the primary instance goes offline. This allows the primary instance to be taken offline immediately for maintenance without causing packet loss, as the standby is already ready to receive and forward traffic.
3Speed
If health probe frequency is reduced to minimize tunnel reconnect time, then switchover speed is improved, but detection precision of gateway health status deteriorates
Solution Approach 1:
The standby gateway instance performs preliminary actions by pre-establishing tunnel connections before switchover is needed. This eliminates the need for frequent health probes during switchover, as the standby is already prepared and can take over immediately when triggered, maintaining both fast switchover and accurate health monitoring.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Techniques are disclosed for managing gateway switchovers. An indication is received that a primary gateway will be switched to a backup gateway. In response to the indication, a response is made to a periodic health probe that a gateway switchover has been initiated. Incoming data traffic is forwarded from the primary gateway to the backup gateway. Subsequent to an elapsed time delay, a response is made to the periodic health probe that the primary gateway will no longer accept incoming data traffic. The time delay may be based at least in part on one or more of a time interval of the periodic poll and a time to effect the gateway switchover. The forwarding of the incoming data traffic from the primary gateway to the backup gateway is terminated.