Subject Re: [IBDI] Classic x Super Server
Author Fabricio Araujo
--Original Message Text---
From: Miguel Angelo Henley
Date: Tue, 4 Apr 2000 12:44:01 -0300

Can any one explain me the difference about Classic Server and Super Server ?

Of course.
To not create one of those messages of 2 pages of paper, let's go in what
make the difference: user conections. Or: what happens when that user connects?
Answers follow:
Classic Architecture(CA).
* A new IB server process is created.
* Every user have it's own process.
* UDF easier to develop (no worry if it is thread-safe or not)
* Individual connections are easier to kill, is only a question of locating the process and kill it
* Classic server eat more computational resources that SuperServer
* Not implement any shared cache for better performance

SuperServer Architecture(SSA):
* Each user have it's own *THREAD* (not a process)
* SSA servers eat much less resources, resulting in the ability handle more connections than CA servers
* Since SSA have a shared cache, it gets better performance
* UDFs not so easy to construct (MUST be threadsafe)
* As far I know, there's no way to kill a individual connection with shutdown the server.

[]s Fabricio