Is OpenDNS doomed?

[mrk]

It's a Saturday morning ¬_¬

 

[Phemo]

Google DNS it is I guess!

Those results fall into margin of error. Changing your DNS server will not affect any latency to any server other than DNS servers so as a result all it can do is speed up the initial lookup when your client tries to connect to a hostname.

Once you've got the corresponding IP address or addresses it's cached anyway. It will not affect your routing in any way.

Edit: Sorry I see you've already responded to another post saying this

 

[Deleted member 77746]

Just changed myself over to google dns, thanks never knew they had one.

 

[rypt]

using google dns as well now instead of easydns

 

[dangerstat]

This seems to resolve URLs much faster here (3 Mobile). Changed and will be keeping with Google's DNS servers for sure

 

[NathanE]

Here are some benchmarks I did using 'namebench'.

These graphs were generated using 30,000 queries which took over 2 hours to complete. So should be pretty comprehensive.

Not sure why but Google doesn't seem to be performing as well as you'd expect

My ISP is Nildram.



Mean Response Duration

Fastest Individual Response Duration

Response Distribution Chart (First 200ms)

Response Distribution Chart (Full)

 

[DJMK4]

Im going to run this on my server in abit. On Be Pro, will post the results when I have them

 

[gurusan]

tbh no ones is going to notice 10ms difference in look-up times. I am just using the google ones because the DNS addresses are so easy to remember

 

[AbsenceJam]

I used DNSBench and they're slower than OpenDNS, Level 3 and some others with less reliability. You're not going to notice a difference but OpenDNS does have SmartCache so you can access sites when their DNS is down. I do like the memorable IP's; although I know L3 and OpenDNS already.

 

[gillywibble]

Namebench gave BT top and my current ISP second. Open DNS was fourth and Google DNS way down in seventh.

 

[Curiosityx]

Here are some benchmarks I did using 'namebench'.

These graphs were generated using 30,000 queries which took over 2 hours to complete. So should be pretty comprehensive.

Not sure why but Google doesn't seem to be performing as well as you'd expect

A very useful tool yes but in order for it to be valid and fair you would need to run then test from hosts connected to many different ISP's with presence in both in the UK and internationally.

Google will always have the advantage when deploying such services as there network consists of well over 400,000 hosts distibuted worldwide, yes it may well be the case that DNS services arnt enabled across them all but the likes of OpenDNS have to make use of collocation services, they cant come close to the footprint that Google has.

More hosts to distribute requests amongst, more responsive the service, latency doesn’t come into it when you talking <30ms.

 

[mrk]

woops