RexxINet maintains a "cache" of web pages and Ftp'ed files on the computer's hard drive (ie, in a particular directory on the hard drive). Any web pages or FTP'ed files that your script (or any other program using RexxINet) asks to receive are stored in this cache (unless you tell RexxINet not to do that when you request some item). As the cache fills up to a certain size limit (as set by the user), then older files are deleted from it.
When you call some RexxINet function to retrieve a web page or file that already happens to be in the cache, then RexxINet fetches the item from the cache (instead of slowly going over the internet to get the same item) if the cached item isn't older than the same item on the internet. RexxINet does this without you even knowing about it, and it happens even if you're in offline mode. RexxINet maintains this cache for the sake of speed. It's faster to pull something off of a local hard drive then get it off the internet.
Persistent cookies are also stored in the cache.
Note: It is possible to tell RexxINet not to fetch a certain item from the cache, even if found there, and instead go get it over the internet. It's also possible to tell RexxINet not to place a copy of some item you fetch into the cache. There are numerous RexxINet functions you can call to enumerate/delete/add items to the cache. But mostly, the cache is there for RexxINet's internal use, to speed things up. You just need to be aware that, if you're in offline mode, then you're restricted to fetching only those items that are already in the cache until such time as you can get connected to the internet.