What FUD. Was this statement from the FAQ not explicit enough? "With Google Public DNS, we collect IP address (only temporarily) and ISP and location information (in permanent logs) for the purpose of making our service faster, better and more secure. Specifically, we use this data to conduct debugging and to analyze abuse phenomena. After 24 hours, we erase any IP information."
Or what about this more detailed explanation on the linked privacy page? [Originally posted, but removed on account of formatting issues]
If you don't know the answer to a question about a competitor's service, the appropriate, ethical thing to say is, "I don't know." When you say, "We do <good thing X>, I don't know if they do," you are by your omission sowing the seeds of fear, uncertainty, and doubt in the audience, for your economic reasons.
They couldn't keep group archives up due to the "maintenance overhead", so they are cost-conscious. They are way beyond being "no evil", so this is hardly a charity for a greater good. There must be the reason.
I doubt that running Google DNS costs a huge amount of money, so the incentive to Google can be pretty subtle and still be worthwhile.
The same reason we've been giving since we released it: to make the Internet faster for people.
> They are way beyond being "no evil", so this is hardly a charity for a greater good.
No, it's because when the Internet works better for people they use it more, and we make more money.
> I can't imagine how tempted they are.
Isn't "friction" what all the startup folks are talking about these days? And how to remove it? It does indeed suck if your business relies on that friction for its business model.
That's more information than anyone should need for Google DNS.
> We don’t persist logs for our users without accounts and configured networks, I’m not sure Google makes the same statement.
Does that mean that both my DNS requests AND my HTTP requests to whatever webserver is intercepting my requests to test.invalid (http://guide.a.id.opendns.com/?url=test.invalid) are not being logged? Does that mean my requests on wifi hubs that are configured with OpenDNS are being logged?
What does it mean to discourage automated DNS lookups? What else do I use it for? dig?
Couldn't one argue that it is a good idea to keep the number of companies that have access to your information low? From that perspective wouldn't it be prudent to use Google for everything?
You and waffle_ss both mention this, can you expand on what problem you're facing. I use OpenDNS because of the filtering abilities and because I found on test that they were marginally faster for me than Google.
Next to never do I see their domain redirect page and whenever I have it's always had the domain I've been after at the top of the page - for example, http://guide.opendns.com/?url=ycambinator.com. Yes it has a couple of text-ads but for the 2-3s you're on the page I can't see that I really have any problem with this at all ... it's way less intrusive than the ads on most websites now.
So what's the issue? Is it really akin to being robbed at gunpoint?
Hmm.. minor correction.. It seems that, with OpenDNS, test.invalid returns NXDOMAIN, test.invali gives me 126.96.36.199. Oh well. When I made that other post I assumed they would be the same.
>they are causing DNS requests to return against-spec responses //
That's part of their service. If there was no way to switch it off then I can understand being annoyed but you can just choose to use your ISP's DNS.
It just appeared to me that both comments concerning this were of the form "ZOMG they has borken my internetz"; could be I read the tone wrong.
So anyway, for the service that OpenDNS are offering is it wrong of them to simplify the situation for users making mistakes entering domain names in their browser?
Of course. But I think that's exclusively the use case that OpenDNS target in their consideration of non-resolving domains.
I think the main point here is persist; OpenDNS probably logs all your requests but they will discard it after x days.
The problem is that any domain lookup requires at least two requests: one to the root servers to find out the domain's nameserver, and one to that nameserver to find out its actual records.
Google has so many users that it's very unlikely that it isn't already in cache, but as a single user of my own resolver, I'd have to pay that penalty for each domain every couple of hours or less (there are some ridiculously low TTLs out there).
Frankly, I think it's worth it.
CDNs using GeoDNS will assume your location is the google DNS servers, and will use the closest edge node. You would likely get better performance with some sites (those using big expensive CDNs) if you pointed to your ISP's recursors (or used unbound to be your own).
: Google's DNS servers are likely anycast multi-homed as well, so it may not be quite as bad as if google only had a couple of centrally located servers. It would still likely skew your closest CDN Edge node a bit.
"I think Google controlling search, the browser, and the network or DNS layer is a dangerous trifecta that the consumer will probably be best served avoiding"
> the interview contains no actual facts
The headline never asserted the article contained facts though? A brief article that contains someone's thoughts is fine by me.
I see nothing disagreeable in the argument that consumers should avoid consolidating all their network activities so those activities route through one advertising corporation.
First page, for those that didn't realise: http://www.forbes.com/sites/eliseackerman/2012/02/25/a-close...
ISP A making their very fast service look dog-slow with terrible DNS (and a terrible wireless router, to boot). ISP B sending mangled responses to some popular requests including facebook.com (which is what made fixing it a priority in my household) and so on...
Warning about consolidation of services would very much be wrong in this situation.