Domains May Disappear After Search 379
Ponca City, We Love You writes "Daily Domainer has a story alleging that there may be a leak that allows domain tasters to intercept, analyze and register your domain ideas in minutes. 'Every time you do a whois search with any service, you run a risk of losing your domain,' says one industry insider. ICANN's Security and Stability Advisory Committee (SSAC ) has not been able to find hard evidence of Domain Name Front Running but they have issued an advisory (pdf) for people to come forward with hard evidence it is happening. Here is how domain name research theft crimes can occur and some tips to avoiding being a victim."
This has been happening a long time (Score:5, Interesting)
My buddy and I even made up names with random letters in a string of 15 or 20, then some porn words stuck on the end ".com".
Sure enough, two days later some squatter had them.
I think the leak is in the registrars themselves. Imagine the money someone could get from the squatters by simply setting up a script to automatically email these queries somewhere.
"Never a more wretched den of scum and villany" describes the whole domain registration process pretty well I think.
Re:This has been happening a long time (Score:1, Interesting)
One would think that in a predatory environment like that, the squatters are doing that to each other already.
Surprised random strings worked.
Don't use Godaddy (Score:3, Interesting)
MD5 lookup as defence (Score:5, Interesting)
Domain tasting is wrong and evil (Score:5, Interesting)
Much of not most of the spam I'm deflecting nowadays seems to come from 'tasted' domains. Or just made up. I almost don't care about the difference.
The last time I read about this, more than a month ago, one snarky idea was to script a tool to randomly taste domains, constantly. If the registrars are forwarding the requests to squatters, they would go crazy with the surge in requests. The squatters would fritter away resources keeping up with these random searches, and eventually the WHOIS functionality of the registrars would have to change. And the script would change, and so on.
I think domain tasting ought to go away, or cost something. $2 for a 14 day taste would wreck the economics, maybe, certainly if random search scripts got going. My server could probably do 100,000 searches a day. I know it can send out 3-4 million spams a weekend, sadly.
Of course, the registrars could block my IP after a while. And blocks of IPs. So we need a Seti@Home-type script that hammers these things out, and let them block every dialup/dsl/cable/sat block. Hehe.
No, it's not devious enough.
Common sense (Score:4, Interesting)
I'm thinking that I'm not liking the direction this is going...
Sniffing, tasting, hmmm, what comes next, digesting? Excreting?
Re:never use the web for such queries (Score:5, Interesting)
Would it help anyone to know who took the domain? I can't seem to get to the article yet.
Re:never use the web for such queries (Score:3, Interesting)
The best protection is to keep the 'window' between testing and registering as short as you can manage, preferably no more than a few *minutes* !
Re:This has been happening a long time (Score:5, Interesting)
So there's the answer to the problem. Bombard the servers with requests for random names. The sleazoids will be forced to either go through the names manually, looking for likely candidates, OR they'll have to register everything...which might tend to get a tad expensive. A script that would hit the whois server with a single randomly generated name every time someone logged into a linux box would probably not put undue hardship on the root servers, but still generate way to many names to feasibly register.
The way to break a scam is to make it expensive to continue. A similar scheme could work for spam. Go through the filtered emails, making a list of URLs. Wait for slow network usage, and do a throttled wget to
it HAS been happening for years. (Score:2, Interesting)
Comment removed (Score:3, Interesting)
Re:MD5 lookup as defence (Score:3, Interesting)
They have the list of the domain names. They only need to calculate a forward MD5 checksum on each domain, and build an index with the MD5 checksum as the key. As new domains are added, checksum them and add them.
Re:This has been happening a long time (Score:4, Interesting)
if a concerted effort were made to cause them to truely jam up the system with this. We could potentially cause them to have a cost. you see...they can taste and taste but realize that there is a bigger fish who is letting them taste his waters.... the registrar that allows tasting.
So... right now, domain squatting is a headache for us, but overall, a minor one, and an even more minor one for the resgitrar. If we could hit them with enough queries, that they truely "taste up" the system... you do two things....
1) You decrease their profit per domain
2) You cause headaches for the registrar as you turn up the volume and jam things up for everyone else
thus... you make their bottom line a small bit worst, and their cost to the tit they are feeding off of go up.
Do it enough and they will either have to stop using whois, or the registrars will stop letting them taste.
Either way, its a win for everyone else. This is totally one of those things where the situation needs to get worst so it can be made better, there is currently just no real pressure on the registrars.
I say.... jam up whois with queries!
-Steve
This is old news (Score:3, Interesting)
Re:This has been happening a long time (Score:3, Interesting)
Re:This has been happening a long time (Score:4, Interesting)
Re:Data mining (Score:4, Interesting)
The obvious disadvantage is that they can't use one registrar to determine that a domain is available and then shop around and use a cheaper registrar to actually buy the domain.
The advantage is that no third party squatter will be able to snipe the domain for themselves - unless of course they use the same registrar.
Re:never use the web for such queries (Score:5, Interesting)
Doing a whois request at a reliable registrar's web-site doesn't go through your ISP's DNS. The larger registrars are probably more trustworthy than your run-of-the-mill ISP. For example, I believe GoDaddy and Network Solutions have stated that they would never provide such information to third parties.
You sure about that? (Score:3, Interesting)
Actually, Bob Parsons (CEO of GoDaddy) has been complaining about "domain tasting" and "domain kiting" for years. Google Bob Parsons domain tasting [google.com] and look at the results. I wouldn't be surprised if it's happening upstream from Godaddy, but I'd be shocked to find Godaddy is in any way willingly facilitating the practise.
What registrar registers a domain for $2? (Score:4, Interesting)
Re:never use the web for such queries (Score:4, Interesting)
The domain wasn't registered when he queried it. But since he didn't buy it right then and there, it WAS registered an hour or so later, by the very site he typed it into.
This has been going on for years, but now the scammers don't even have to rely on roommate stupidity.
Re:never use the web for such queries (Score:3, Interesting)
Re:its actually pretty common (Score:3, Interesting)
Re:Data mining (Score:4, Interesting)
Now, the squatters COULD start developing a list of IP addresses that are doing lookups, and filtering them out of their results. Of course, this would be all right as it would mean you were protected from someone sneaking in and squatting the name you looked up. Even if the squatters filtered on both IP address AND multiple hits, this could be resolved by allowing real name lookups to be submitted into the random name lookup web site. Then if you wanted to lookup ihatedomainnamesquatters.com, not only you but everyone else that has been looking up random names, will look up ihatedomainnamesquatters.com also. It would be virtually impossible to tell the difference between real interest, and fake.
Plus, if you wanted to both fund the site AND be ironic, you could put advertising on the web page.
Re:never use the web for such queries (Score:1, Interesting)
Keep on believing that, but both of them either sell that info or buy the domains directly (through some shell companies) or they have malicious employees selling that data.
I've done whois lookups at both of them using some pretty obscure domain names, only to have the domains purchased by someone 2 days later. It appears who ever bought them was just tasting them because the domains because available again few weeks later. But it does show that someone is sharing/selling data.
I've never trusted Network Solutions, I use to trust GoDaddy, but after that I've switched everything over to PairNIC. The one and only web host I trust running the now one and only registrar I trust.
Re:Data mining (Score:4, Interesting)
The stated reason for allowing retraction of registrations is to allow mistakes to be corrected. But with domains costing just a few dollars to register for a year, how much harm is done by making the customer pay for such mistakes? Answer - none at all. Meanwhile unscrupulous domain tasters are registering, and then returning, millions of domains a day for free.
The DNS marketplace has probably the most widespread corruption of any economy in the world today.
Easier solution (Score:5, Interesting)
Re:never use the web for such queries (Score:4, Interesting)
Just to present a counterpoint: a couple of years ago, the opposite happened to me. I registered a domain name based on the name of my character in an online game. It was certainly an unusual name that I had never run into.
A few days later, I got a somewhat angry email from someone wanting to know why I had taken that name, because it was their surname, and they had planned on registering it. Once I explained the situation the guy calmed down and all was well.
But the moral is that it is quite possible that someone, completely innocently, took the domain you were researching, within a day or so you doing it, because that's exactly what happened with my domain. In my case, I just got lucky... 2 days later, the domain would have been gone.
Re:This has been happening a long time (Score:4, Interesting)
These are the steps that should be taken:
The nice thing about the scheme is that squatters could be aware of and even secretly participate in it and it would still work. They'd have no better chance of identifying legitimate queries from random queries. And they can't exactly poison random data.
Re:Data mining (Score:4, Interesting)
THIS is one of the things they are trying to prevent.
Re:we got tasted.. (Score:3, Interesting)