I've got a question regarding the user of Socket communications via KixForms.
I'm thinking about a service that hands out configuration settings for an app. When the app starts up, and every 1-4 hours thereafter, it will check in. It will make 4 specific data requests, which the server will look up (initially in an INI but later in a SQL table) and return to the client. The client will be on every server, from 20 to around 500 servers depending on the environment.
If I have the server listening on, say, port 4500, should I have the server service respond directly, or should I spawn a second, temporary listener and have the server simply reply with the child's port number? The client would then communicate with the child listener on the specified port, and the child would terminate after the final query or 60 seconds. Basically, I'm wondering if a single listener could process simultaneous requests from multiple clients.
FYI - the query is generally in the format "HOST:REQUEST_ID" and the response, based on the Request_ID, is under 1K. Occasionally (once monthly), the client would send about 200 bytes in addition to the host ID and Request ID.
Thanks for your thoughts..
Glenn
_________________________
Actually I
am a Rocket Scientist!
