Re: Stale entries in dnscache's cache structure

On Wed, Mar 10, 2010 at 4:17 PM, Matthew Dempsky <[email protected]> wrote:
> I think it would be interesting to see comparisons on how dnscache
> behaves under different cache implementation strategies (e.g.,
> memcached's probably has very similar requirements, since it's a fixed
> memory key/value store with expiration times).  The existing approach
> is certainly remarkably simple and demands very little CPU work for
> reads and writes while also keeping memory overhead to a minimum.

A while back, I was experimenting with replacing the existing cache
with SQLite as a way to have multiple dnscache instances share the
same cache without having to write any synchronization / locking
logic.  Now that you mention it, memcached would be even better...