On Feb 27, 2008, at 11:09 AM, William Herrin wrote:
It's not just state, it's churn. It's 18 minutes that you have to keep receiving and asking to receive map updates for the destination even though there's no associated user-level traffic. That or the ITR has to do some sort of probability assessment as to whether the expired data is still valid so that it can use the expired data for the next NTP packet while it goes out and refreshes the cache. I've explored this possibility a little with TRRP. It gets hairy fast. You can't simply extend the cache validity of course. 18 minutes is too long for multihoming, let alone mobility.
Bill,Your assumption here is that all cache entries need to have equal validity. This is not necessary; it's trivial for the cache entry to have an associated TTL that tells the ITR exactly how long to retain the cache entry. I'll happily stipulate that interesting NTP servers will need a cache entry TTL of more than 18 minutes. I hope that's not an issue to you.
For mobile hosts (i.e., laptops), you crank the TTL down. For fixed hosts, you crank the TTL up. Straightforward and simple.
Tony -- to unsubscribe send a message to rrg-request@psg.com with the word 'unsubscribe' in a single line as the message text body. archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg