Suresh Krishnaswamy: Re: [dnssec-deployment] Future applications?
mayer at gis.net
Mon Jan 14 20:07:13 UTC 2008
Paul Vixie wrote:
> what asynchronous dns api do folks want in BIND?
I think you know NTP's needs. In this case it's mainly a need to call
getaddrinfo() and pick up the results later. Adding an DNSSEC
authentication option to require that the data has been validated would
be useful along with a mess of error codes to tell the application the
result of the validation. Is it possible to have a different validation
result for each record returned?
In the more general case there's also needs for applications to fetch
other record types, MX and SRV are the currently most likely ones.
There is also a need sometimes to return the TTL of each record.
getaddrinfo() does not do that, yet I saw an RFC (RFC 3435) recently
that was expecting the recipient to not only use the addresses returned
but also their TTL so that they do not get used after they expire. I'm
not sure how they were going to implement that unless they made their
More information about the bind-workers