Link restoration method based on switch and controller, controller and switch

Abstract

The invention provides a link restoration method based on a switch and a controller, the controller and the switch. The link restoration method includes the following steps: S1. establishing a first link of the switch and the controller, and establishing a TCP session between a port of the switch and the controller for performing Openflow message interaction; S2. sending port state information tothe controller by the switch; S3. performing retransmission of the TCP session according to a manner that a time interval is gradually increased, if the first link of the switch and the controller isinterrupted, and the sending of the port state information is timed out; S4. determining a link restoration strategy according to a relationship between a retransmission timeout interval and a presetthreshold, wherein the strategy includes a timeout waiting strategy and a fast reconnection strategy. According to the retransmission timeout interval, a corresponding link restoration strategy is determined, and link restoration is performed by adopting combination of "timeout waiting" and "fast reconnection", thus reducing the time interval between link convergence completion and the link restoration of the controller and the switch link after a link fault of the switch occurs.

Description

technical field [0001] The invention relates to the technical field of network communication, in particular to a link restoration method based on a switch and a controller, a controller and a switch. Background technique [0002] In the prior art, the controller and the switch are connected by means of TCP. TCP (Transmission Control Protocol) is a connection-oriented, reliable, and byte-stream-based transport layer communication protocol. The prior art uses TCP timeout retransmission The mechanism is used to reconnect, and the time between reconnections adopts an "exponential backoff" algorithm. [0003] The TCP retransmission timeout interval adopts an "exponential back-off" algorithm, which means that when repeated conflicts are encountered, the station will repeat the transmission, but after each conflict, the average value of the delay will be doubled until it reaches A maximum interval time; for example, the first interval is 0.2 seconds, the next interval is 0.4 secon...

AI Technical Summary

Problems solved by technology

[0007] 3. The network link with a time stamp of 18.22 seconds fails, and the message cannot be transmitted normally;

[0016] In the above example, there is a difference of 6.8 seconds between network recovery and controller connection recovery (ie steps 9 and 10)

Images