Re: [dna] Review of draft-ietf-dna-simple-11

["Bernard Aboba" <[e-mail]>]

Thomas Narten said:

"Note, as a comparison point, a stack that does not implement DNA. When
the node encounters a "link up" event, that presumably also implies a
"link down" event preceding the "link up" event, and the default
course of action should be to throw away all the old configuration
information and then start the configuration process from scratch."

[BA] Reception of a "link up" event doesn't necessarily imply a previous
"link down" event, nor should reception of either a "link down" or a 
"link up" event necessarily result in throwing away old configuration
Information. 

As Ted noted, a "Link Up" event can be received as a result of jitter
on wireless interfaces, in which case the IP address might have remained
unchanged.  A "Link Down" indication can be followed by a 
"Link Up" for the same network, so that discarding configuration
Information in this case may also not be a good idea. 

See RFC 4907 for more discussion on this.