Both the Classic OpenEdge AppServer and the Progress Application Server for OpenEdge are scalable—they can be configured to optimize performance as the workload increases. Note, the differences in how you scale each type of server in the following areas:
Adding and trimming agents — The first step in scaling a server is to adjust the number of agents in order to maximize the server's response time under a given workload.
In both the Classic OpenEdge AppServer and PAS for OpenEdge, you can edit a properties file or use OEE/OEM to adjust the number of agents running on a server. However, PAS for OpenEdge also allows you to modify the number of sessions that an agent supports. An agent in the Classic AppServer is a single session agent, whereas PAS for OpenEdge agents are multi-session agents.
Load balancing — The second step in scaling is load balancing. Load balancing basically involves adding more instances of the server and configuring an environment where the workload is share among those instances.
With the Classic OpenEdge AppServer, you can register multiple AppServer instances that support the same application services using an OpenEdge NameServer. You assign a weight factor to determine how the NameServer balances the load (client requests) among instances.
You can create multiple server instances, but there is no NameServer in PAS for OpenEdge. Instead of a NameServer, load balancing is achieved in PAS for OE through an Apache extension module or through third party applications. For example, see:
Adding and trimming agents — The first step in scaling a server is to adjust the number of agents in order to maximize the server's response time under a given workload.