I am in the process of creating an online implementation of a very processor intensive application. The model I am contemplating will work something like this
[list][*]
The user logs in through a web browser and requests an online session
The server logs this request to a database.
The request is picked up by a [B]monitor/marshal
The lightweight executable, [B]the runner[B]originating client
The webserver serves up a basic DHTML/AJAX web page to the originating client and then steps out of the picutre.
From that point on until the online session is terminated all further interaction involves peer-to-peer communication between the [B]client[B]runner
Once the online session is terminated the [B]runner[B]marshal
The question I have is this – will I run foul of any client side firewall/browser security requirements by getting the browser to perform http on a non standard port and, possibly, running on another computer (different IP) than the webserver to which the original online service request was made.
Not sure that I am expressing myself with adequate clarity here but I would much appreciate any help.