How to Compare Hosted DNS Providers (with Data)

dns.learnhub.com

36 points·jpg··5 comments

DNS response time can be a huge issue. The type of software you use can also make a difference. I noticed a significant increase in the speed of lookups when I switched from using BIND9 to djbdns.

I haven't heard of djbdns... did you measure performance before and after? How did you measure it?

djbdns id from Dan Bernstein - http://cr.yp.to/djbdns.html

No, I didn't do any measurements other than subjective observations. This was on my LAN and using BIND9 I would see noticeable delays between the 'looking up' and 'connecting to' messages in firefox. After switching to djbdns you hardly have time to read the 'looking up' before it starts the connection.

I'm curious if Pingdom's DNS testing is sufficient to draw performance conclusions about these services. I ask because the summary results table is curious: it only lists the first NS record for each hostname. (NB: luilui.com does not see to use netriplex currently). If 14 worldwide testing nodes are pegged to try one nameserver and that nameserver is not an anycast node, it seems to me this test is not accurate. If someone knows what Pingdom is doing behind the scenes -- or if this is just a data display issue -- I'd love to know.

I only ran the test against the first name server listed. It would have been more complete to test again all/several, but I thought this would be sufficient.

LuiLui.com has changed its name servers since I ran the test. I just used the first domain that I found that used them.

Regarding Anycast, all 3 providers advertise as such, so I assumed that all their name servers would use it

My reason for publishing this is that I couldn't find any data comparing providers, or even any good ways to compare them.

Reply on news.ycombinator.com