Hello! Task: Choose best hosting for site (analize other sites) Have: 10000 urls Need: Check IP, country, hoster Check IP and country not very hard, but how I can know hoster? How I can 100% check this? For example, I can grab info from here: But for guest I have very limited access... CURL I don't know now... Second trouble, for different zones used different WHOIS-Services and info was different to... I think I need: Name Server, IP Address, IP Location (maybe Reverse IP), and check Whois Server, for "info style". But see to example: Whois Server: whois.enom.com - ОК IP Address: 38.108.68.98 - ОК Name Server: NS5.VEXXHOST.COM - ОК Name Server: NS6.VEXXHOST.COM - ОК IP Location: Canada - Psinet Inc - ??? I can check hoster for NS, but I can use other NS, not my hoster, and I get trash, not true info... Psinet Inc - it's provider of my hoster... next level... Who know more simple and 100% method? Thanks for replies. Sorry for my bad english!
It's not easy; there are so many different ways that this information is recorded. And there is definitely no 100% method other than doing it by hand. The most reliable way to figure out the host is looking up who owns the IP blocks, by connecting via whois protocol to ARIN/APNIC/LACNIC/AFRINIC. I don't think any of these are going to let you make thousands of connections in a day, so you'll have to distribute your task across multiple source IPs or multiple days. You can find a free geolocation database here: http://blogama.org/node/60 - like all geolocation databases (free or paid) it's not very accurate, but better than nothing.
1. I can write different regexp for ARIN/APNIC/LACNIC/AFRINIC. I have GeoIP (from MaxMind)... I don't want do this faster... But I don't know HOW I can 100% check hoster... NS Server? IP range? or How?
Like I said, the only way to do it 100% is by hand. There are too many inconsistencies in the way that various providers and operators report things, and too many complexities in the way hosting is arranged. Remember, "hosting" isn't a black-and-white concept. What if I have round-robin DNS with servers at multiple hosting outfits? What about organisations with their own assigned portable IP space that self-host? Do you consider them the "hoster", or their upstream? What if they have multiple upstreams? What about colocation centers that delegate IP space to their clients who in turn re-sell to customers? What if those customers further re-sell to other customers? What about the fact that some of these also subdelegate whois but others do not? All of these arrangements are extremely common and together account for a large share of the sites out there on the web. I doubt you would ever be able to come up with an automated system capable of accurately sifting through them all. Even a human would have a hard time unless they have a lot of experience in the business. A quick example: I have a bunch of servers at Rackspace which I use to host client sites; if you follow the whois chain, some of them end point to me and others point to Rackspace. For the IPs I requested one-by-one, they tend to be listed as Rackspace and it stops there. For the ones I requested in larger blocks, Rackspace mostly (but not always) publishes whois data listing me as the owner.
Bad news I have some sites sorted by my category. I want to create site in this category but some hoster in TOS not allow some content. If I can check this sites (maybe not 100%... 70% is good!) I can find loyalty hosting for my site. For example, my hoster not support pr0n, war3z and anime... but if I want to create anime site I must find another hoster... read TOS in all hosters + call to support and ask him... not good variant. If my script can detect right 70% sites I can choose hoster for my new site!