Hi Paul On Thu, Jun 25, 2020 at 02:29:03PM -0400, Paul Wouters wrote: > On Thu, 25 Jun 2020, Mukund Sivaraman wrote: > > > For whoever is interested, this is a description of a pattern of queries > > noticed at busy public resolvers that has led to issues in at least 4 > > different sites in the last 2 months. > > > > The current revision is a work in progress. We are still developing some > > mitigations for NIOS, and some more introductory text also has to be > > added. > > I would add a more explicit section on using prefetching of frequently > asked queries, which mitigates (eliminates) the period when an answer is not > available in the cache.
The resolvers in question had prefetching functionality enabled. It
didn't help mitigate the problem. These are very busy public resolvers,
and their clients for the most part are static participants (in other
words, clients don't join and leave the group in large numbers in the
TTL interval). So the clients align themselves into herds at the expiry
time, and they show up in a spike that is instantaneous. The frequency
of these spikes was observed to be equal to the TTL of the
answer. Prefetching does not happen as there's often no lone client that
comes in a few seconds before the answer expires. The answer expires and
the spike causes nuisance.
Your suggestion to mention prefetching is good though and I will add
notes on what happens with prefetching.
Mukund
signature.asc
Description: PGP signature
_______________________________________________ DNSOP mailing list [email protected] https://www.ietf.org/mailman/listinfo/dnsop
