How are ajax/websocket domain searching services so fast?

I have been using ajaxwhois for years, and there are many of these services for finding domain names now.

The fastest seem to use some sort of websocket to connect the interface to the data; but I still don't quite get how it is so lightning fast.

Simply doing a whois whatever.com from a shell has a noticeable delay. Are thes开发者_开发知识库e services really just making a bunch of direct requests (in parallel) using whois and parsing the data, or is there some other method that is escaping me?

edit: I should be clear in that I am not confused/curious about the speediness of ajax/sockets; even assuming the server->interface delay is 0ms I don't understand how the queries themselves are being made so quickly.

edit2: links

http://ajaxwhois.com/

http://www.domainzomg.com/

---

My guess is the server caches the results, and probably refreshes them server-side every x hours. So all the AJAX script does, is get the cached result. You can test this by entering a common domain name like mathematics.com, where the search results will be very fast, and then an uncommon domain name like howareajaxwebsocketdomainsearchingservicessofast.com which will show a noticeable delay, the same showed by a shell command.

---

You could also do ns lookup as a part of the check. Other ways is also possible. For example many registrars has free servises that only respond if the domain is free for regitration that is mutch faster that parsing whois. If you have access to epp you could also check in the registration system.. so there is 4 ways to check. It all dep on what TLD and