[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Multihoming an end user



At 01:37 PM 9/24/05 -0400, dan@thecsl.org wrote:
>Is it common for remote caches to ignore TTL and negative TTL?
>
>With a tiny (< 100 connections per day) user base, you can pretty 
>comfortably set DNS TTL and negative TTL to 60 seconds, even with 
>caching 3 levels deep, it will take 3 minutes for the right IP# to be used.

Well setting a very small TTL is what u would have to do but it's not
something that I would trust for any kind of "snap" change.  As for the
random BIND configs on the internet, there's no telling.

>Also, if as was the case with the OP, ISP1# cuts service off explicitly 
>and deliberately for political reasons, they are also likely to suddenly 
>stop routing.
>How would such a cut-off effect the whole 
>ARIN/ASN/BGP/magic-beyond-my-humble-understanding thing?

In that case everything would fail over to ISP2 since there is a valid world
known route to u through that ISP.  Now theoretically if u got ur block from
ISP1 they could take back their IP block and kill ur ASN routability.
Getting an independent block is the way to keep that from happening.  No
matter what either ISP does u still control the routing.  Of course u need
some ISP to push that route out to the world.


Another question for the original poster:  How is it that u have two DSL
lines from two providers?  Usually one provider controls the local loop, and
in that case both DSL lines would be going out over the *same* CO, DSLAM,
ATM, etc until it branched out to the ISP's network.






--
REMEMBER THE WORLD TRADE CENTER         ---=< WTC 911 >=--
"...ne cede malis"

00000100



Reply to: