Data collection

We have an mysql database with about 20000 urls We would like to get an spider who spiders all url and puts all found words on the url in a mysql database. Found words must be unique in combination with url So a word can only be stored one time behind an url. Good example: url word [login to view URL] software [login to view URL] and [login to view URL] software [login to view URL] software Bad example [login to view URL] software [login to view URL] and [login to view URL] software wrong [login to view URL] and wrong [login to view URL] software [login to view URL] software Even if the word whas found on [login to view URL] it should not store this data The created database will have only six fields Url,word,date,time,url where data whas found example off data [login to view URL],software,12-12-2006,10:10:10,[login to view URL] Where date and time are the date and time that the spider got the word from the website. This spider should run continously and if finnished with all url start all over again with the first url. iff on any moment extra urls are put in the database then offcourse the new urls have a prioritie over the allready indexed urls. The spider should stay at the domain : So if the start url = [login to view URL] it should also go to [login to view URL] and also to [login to view URL] but not follow a link to [login to view URL] for this is a complete different site. The spider should be using as much treads as possible with te available bandwith. Because off a bad experiance in the past we will pay the agreed price after we see the program running on your server and see the data collected from like 50 sample url.