[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: More REAP cases
Erik Nordmark wrote:
This case is already in the draft, but it makes sense to add the timer
and
more packets getting lost to show what happens:
Looks good, thanks!
This shows when the failure for the current locator pair is in the other
direction.
Peer A Peer B
| |
| (A1,B1) payload packet |
|-------------------------------------------->|
| |
| (B1,A1) payload packet |
| /-----------------------------------------| Time T1
| | Path B1->A1
| | is now
| | broken
| (B1,A1) payload packet |
| /-----------------------------------------|
| |
| (B1,A1) payload packet |
| /-----------------------------------------|
| |
| | 10 seconds after
time
| | T1, sends a com-
| (B1,A1) REAP Event id=p, | plaint that
| iseeyou=no | it is not rec-
| /-----------------------------------------| eiving anything
| |
| (B2,A2) REAP Event id=q, |
| iseeyou=no | Next try different
|<--------------------------------------------| locator pair
| |
Add "this one gets through"
| (A2, B2) REAP Event id=r, |
| iseeyou=yes |
| payload reception rep |
| event reception rep(q)| This one gets
|-------------------------------------------->| through
s/This one gets through/now A responds/
| |
| |
| | B now knows
| | that A has no
| (B2,A2) REAP Event id=s, | problem to receive
| iseeyou=yes, | its packets and
| event reception rep(r)| This one gets
|<--------------------------------------------| that A has found
| | a new path to B
| |
| (A2,B2) payload packet |
|-------------------------------------------->| Payload packets
| | flow again
| (B2,A2) payload packet |
|<--------------------------------------------|
Good case too, some nits above.
This case we have even less luck. The response to the REAP event
doesn't make it in the reverse direction, so both ends end up exploring
indepedently.
Peer A Peer B
| |
| (A1,B1) payload packet |
|-------------------------------------------->|
| |
| (B1,A1) payload packet |
| /-----------------------------------------| Time T1
| | Path B1->A1
| | is now
| | broken
| (B1,A1) payload packet |
| /-----------------------------------------|
| |
| (B1,A1) payload packet |
| /-----------------------------------------|
| |
| | 10 seconds after
time
| | T1, sends a com-
| (B1,A1) REAP Event id=p, | plaint that
| iseeyou=no | it is not rec-
| /-----------------------------------------| eiving anything
| |
| (B2,A2) REAP Event id=q, |
| iseeyou=no | Next try different
|<--------------------------------------------| locator pair
| |
A know needs to
start exploring
| (A2, B2) REAP Event id=r, |
| iseeyou=yes |
| payload reception rep |
| event reception rep(q)|
|--------------------------------------/ | Doesn't make it
| |
| (A1, B1) REAP Event id=s, |
| iseeyou=yes |
| payload reception rep |
| event reception rep(q)| This one gets
|-------------------------------------------->| through
| |
| |
| | B now knows
| | that A has no
| (B2,A2) REAP Event id=t, | problem to receive
| iseeyou=yes, | its packets and
| event reception rep(r)| This one gets
|<--------------------------------------------| that A has found
| | a new path to B
| |
| (A1,B1) payload packet |
|-------------------------------------------->| Payload packets
| | flow again
| (B2,A2) payload packet |
|<--------------------------------------------|
Yes.
I have added these to the draft.
--Jari