In a previous episode Stephen R. van den Berg said...
::
::
:: Now, the 30% packet loss on the two unreliable links is such that, if I
:: were to transmit a packet on both links simultaneously, it evens out to
:: almost 0% packet loss (the packet will almost always go through on either
:: link, if not on both).
this intrigues me.. if the links were independent of each other, you'd
expect 9% loss under the duplicated scheme.. and that's a number that
will still give TCP fits.. of course, there's no reason to think that
two wireless links were independent, but I'd expect to see a positive
correlation somehow.. e.g. a weather condition would cause both links
to drop at the same time.. but instead you seem to have a condition
where a failure on one nearly guarantees a success on the other.. I
don't know a heck of a lot about wireless properties, but I'm unable
to hypothesize why that might be? (assuming the links have the same
30% drop rate when the other link is idle..)
got any guesses or educational explanations to share?
:: b. This means that I'd like to check on the Linux box, if I am about to
:: transmit an ACK or if I'm about to transmit a small packet (say <64 bytes).
:: If yes, duplicate the packet and route it to *two* routers on the
:: local ethernet simultaneously.
This scares me some.. it seems to make more sense to bond the channels
somehow and treat them as one link so that the duplicate packet can be
stripped on the other end instead of propogated on.. we've already got
enough duplicate IP packets running around the network from flapping
SONET implmentations..
-P
-
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to [EMAIL PROTECTED]
