M
Mark
We are building a public web application that calls a web service on an
internal box. The web service runs for 5-10 minutes, does some heavy
processing, and writes to a database when it is complete. The current design
calls for having the web application's thread sleep for 10 seconds, then
check the database to see if the web service is complete. If the database
claims the web service has completed, it then displays the results from the
database.
There are a number of ways to solve this problem, but I'm wondering what the
true impact of having the thread sleep on overall performance and
availability of our production web server that hosts many other
applications. Obiviously the thread isn't doing anything, but it still is
tying up a thread. Is this a big deal? Is there a way to quantify or
measure its impact and/or determine how far it would scale?
Thanks in advance.
Mark
internal box. The web service runs for 5-10 minutes, does some heavy
processing, and writes to a database when it is complete. The current design
calls for having the web application's thread sleep for 10 seconds, then
check the database to see if the web service is complete. If the database
claims the web service has completed, it then displays the results from the
database.
There are a number of ways to solve this problem, but I'm wondering what the
true impact of having the thread sleep on overall performance and
availability of our production web server that hosts many other
applications. Obiviously the thread isn't doing anything, but it still is
tying up a thread. Is this a big deal? Is there a way to quantify or
measure its impact and/or determine how far it would scale?
Thanks in advance.
Mark