You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The type = fastest group is great, but maybe we can make it even greaterder :-).
Make it more sticky (as an option), where after a while firing queries to all resolvers, get more sticky to the fastest one more, decrease the number of parallel queries and test "fastest" over time (and decrease frequency as well) to make sure we stay "fast".
This will bring down the load and number of queries and potentially more servers can be used in the pool for redundancy reasons. besides it being super-cool to have ;-).
The text was updated successfully, but these errors were encountered:
Interesting idea. Quite like how that reduces the overhead by not sending multiple queries. It differs quite a bit from the current implementation of fastest, so perhaps it might be best to come up with a new entity for it, rather than overloading the existing one.
The
type = fastest
group is great, but maybe we can make it even greaterder :-).Make it more sticky (as an option), where after a while firing queries to all resolvers, get more sticky to the fastest one more, decrease the number of parallel queries and test "fastest" over time (and decrease frequency as well) to make sure we stay "fast".
It is called Smoothed Round Trip Time (SRTT), see here: https://ns1.com/blog/srtt-and-recursive-dns-resolvers
This will bring down the load and number of queries and potentially more servers can be used in the pool for redundancy reasons. besides it being super-cool to have ;-).
The text was updated successfully, but these errors were encountered: