[dnsdist] Intrigued by c-hashing --- anyone using it?

Alexander Perlis aperlis at math.lsu.edu
Mon Mar 27 17:43:50 UTC 2023

I'm intrigued by "c-hashing". The essential idea seems to be that
different clients requesting a particular CNAME should end up at the
same recursive resolver, reducing infrastructure cache size
requirements and reducing the upstream load towards root name servers
and TLD name servers. Seems like a really awesome feature. Thank you
to the folks who thought of and implemented it!

Is it known whether any large institutions (universities or
enterprises) are using "c-hashing" in a production environment? Any
lessons learned, and/or did you also try other load balancing
algorithms and end up preferring c-hashing?


More information about the dnsdist mailing list