DNS Benchmark N66U

[CaptainSTX]

When I run the latest version of DNS Benchmark it reports that I only have one DNS server 192.168.1.1 (Local Network Nameserver ).

I get the same result regardless if I have selected connect to DNS automatically (Yes) or (No) and then enter the IP addresses for two DNS servers.

I know the two servers work as they both show up in the Benchmark report with statistics in both cases. They just don't show up in the report as DNS servers selected by me and used by my system when I am connected using my ASUS N66U running Merlin's V24

I have saved the settings and rebooted the router several times.

If I connect to the Internet using my E3000 running Tomato and then run Benchmark I get the three DNS servers that I have selected reported correctly.

Have I missed something? Is it possible to get the answer using Benchmark when run behind the N66U?

 

[RMerlin]

The router acts as a DNS proxy. That's why your clients only have your router's IP as DNS, since the router takes care of forwarding DNS requests to whichever DNS servers is configured on the router.

And I'll save you some time by telling you that DNS benchmarks are half useless. Check these forums for why, it has been explained many times already. The short version: saving 10-20ms on a DNS request that will get cached afterward anyway but ending up with a resolved IP that's across the country instead of being closer to your ISP will result in slower download performance.

 

[CaptainSTX]

Thank you for your explanation. I now know what I am seeing on the ASUS is the way it is. As you said for most people having the information doesn't make much difference,

In my particular situation however having the information that Benchmark can provide can be useful.

I am located on a Caribbean island 1,250 miles from Miami.

I use a commercial OpenVPN server in Miami for much of the traffic that I want to be more secure.

If I use the local DNS server provided by my local ISP on the Caribbean island the request for a name resolution has to come back from Miami to the island then the result of the lookup has to be sent back Miami.

By using a DNS server located in the Continental USA closer to Miami I can consistently cut the response time almost in half. Benchmark makes it possible to seek out these servers. I also connect through Stockholm and London using my VPN.

While the local DNS servers on my local ISP can resolve a cached name on average in .007 .010 seconds if they are used to resolve a name when connected through Miami the time increases to .060 -.063 seconds.

Using a DNS server nearer Miami results in an average tome of .036 seconds.

I also like to know who my DNS providers are if nothing else to provide some redundancy so it OpenDNS is number one you can set Google or whomever for number two.

Let me know if there is a fallacy in my thinking.

Thanks for your outstanding software and prompt response to my post

 

[RMerlin]

I see your case.

Ideally, you'll want DNS servers that are near the exit point of your OpenVPN tunnel. That way, geolocalisation will still work to your best advantage. For example, if your tunnel endpoint is in London, using a DNS server from a London-based ISP will ensure that when you issue a lookup for, say, Youtube, you will be returned the IP of a Youtube server located near to London, rather than one in the US.

Unfortunately, a DNS benchmark won't tell the whole tale. If, for example, the Google DNS servers return the answers the fastest, but yet give you the IP of a Youtube server too far away from your tunnel endpoint, your streaming experience will deteriorate.

So yes, using DNS servers from your ISP will definitely not work out for the best if you always use the tunnel. But to find out the best DNS servers to use you don't want to rely on a benchmark, but on the physical location of the DNS servers you wish to use, which must be close to the tunnel's endpoint.