[dns-operations] More Aggressive prefetch for popular names

[Paul Hoffman]

On 15 Apr 2019, at 14:15, Peter van Dijk wrote:

> Hello Paul,
>
> On 11 Apr 2019, at 15:54, Paul Hoffman wrote:
>
>>> "only increasing the authoritative load in the less common cases."
>>> That's where it can get tricky. In theory, yes, but in practice it's
>>> hard to estimate the aggregate effects from such policy -- it may as
>>> well lead to some unintended collateral damage -- and IMO we need 
>>> more
>>> studies on this.
>>
>> Fully agree. This could even be aided by some resolver developers 
>> adding a debug feature.
>
> What would that debugging feature look like?

As a first guess:

In a cache dump or other way of getting resolver stats, you could report 
the ratio of names that had a prefetch that have that data hit in the 
cache before the TTL expires (0% bad, 100% best). For software that lets 
the operator tune the prefetch, this ratio could help in the tuning. For 
software that uses off/on setting, if that ratio is less than 50%, the 
operator might consider turning off prefetching.

Actual operators of customer-facing resolvers (that's not me!) might 
suggest here other debugging that might aid them in deciding how to tune 
or turn off/on.

--Paul Hoffman