Measuring BGP Update propagation time
As the standard ping, the bgp-ping application works by sending and receiving probe messages. A sender embeds a probe in a BGP Update message and sends it to its BGP neighbor. We encode the probe as a custom BGP optional transitive attribute, which allows it to propagate transparently throughout the network until it reaches one of the bgp-ping responders.
A bgp-ping responder recognizes this custom transitive attribute and reflects it back to the sender. This gives the sender a measure of BGP Update propagation delay within the network, which is then reported to an external metric consumer or printed on a screen.
Since the bgp-ping application needs to inter-operate with real BGP stacks, at the very least it has to implement the initial exchange of Open messages to negotiate the BGP session capabilities, followed by the periodic exchange of Keepalive messages. We also need to do the following:
- Send BGP Update messages triggered by...
