I am currently testing a few DNS servers with dnseval. One suggestion: Since it tests basically the "cache hit" rate it makes a huge difference whether the requested hostname is already in the cache of the DNS resolver or not. If it is in the cache, all responses are fast. If it is not in the cache, the very first query is slow while all further queries are fast.
Hence my idea: Could you send a first query to the resolver (in order to have the hostname in the cache), wait a few moments (let's say, 1 second) and then start the actual dnseval test of the servers? To my mind this would give better comparable results.
Hi Babak,
I am currently testing a few DNS servers with dnseval. One suggestion: Since it tests basically the "cache hit" rate it makes a huge difference whether the requested hostname is already in the cache of the DNS resolver or not. If it is in the cache, all responses are fast. If it is not in the cache, the very first query is slow while all further queries are fast. Hence my idea: Could you send a first query to the resolver (in order to have the hostname in the cache), wait a few moments (let's say, 1 second) and then start the actual dnseval test of the servers? To my mind this would give better comparable results.
Ciao, Johannes