Direction needed for constant checking application
I need some direction on how to achieve our goal.
We have an application that will be deployed on hundreds and potentially just over a thousand computers that are not on the same network. This application already automatically updates itself by calling our webservice and checking for a newer version so it "calls home" to an extent at this time.
The goals of this application have evolved (go figure) and now management would like to be able to "ask" this application to perform certain functions on each computer. At the moment, this is as simple as asking for the currently logged in user account.
What they would like is to right click on a computer from their management interface and select "show current user". The application on the computer would then respond with this information.
So, the question is how to make the workstation application stay in touch frequently enough to give management the information they want without waiting very long yet also not overload bandwidth with unnecessary traffic.
Should I have the application just constantly call a webservice every X number of seconds and see if there is any "work" to do? Or, should I somehow keep a connection open? Or some other technique?
Thanks for any advice!
Re: Direction needed for constant checking application
Either your workstation should poll the webservice periodically or be constantly online (like ICQ, Jabber, etc). There isn't any other option really. In order to receive a task the information has to be transmitted from the server to a client.
By the way, you can use jabber protocol for this.
Re: Direction needed for constant checking application
I agree, but will I be generating too much traffic to be polling from a thousand computers every X number of seconds? Let's say X was 15. Every 15 seconds, a HTTPS web request and response (as small as I can make it) will be sent out.
Not sure how much traffic that would be, but seems like it is potentially quite a bit.
Wasn't sure if that was the best route or if I should somehow have "push" technology where my server would tell the computer they had something to do.
Re: Direction needed for constant checking application
Quote:
Originally Posted by
zookie
I agree, but will I be generating too much traffic to be polling from a thousand computers every X number of seconds? Let's say X was 15. Every 15 seconds, a HTTPS web request and response (as small as I can make it) will be sent out.
Not sure how much traffic that would be, but seems like it is potentially quite a bit.
Wasn't sure if that was the best route or if I should somehow have "push" technology where my server would tell the computer they had something to do.
Then the connection should originate from the server. This approach requires keeping track of online and offline clients and ensuring that the client is listens for communication. Still I think that you should use something like ICQ network.
