TCP/IP is the standard communication protocol used by Argent.

And there are a lot of things you can do to optimize the performance; here are some of the more important ones.

TCP Send And Receive Timeout

Setting the right timeout value for TCP send and receive operation is essential to get good responsiveness for both client and server.

By default, TCP blocking sockets wait infinitely when sending or receiving data. It breaks out only if data is sent, received or connection is dropped. This can lead to a deadlock condition when the sender sends partial data and the receiver gets partial data, and they both wait for the other side to proceed.

Setting the send and receive timeout can prevent this deadlock condition completely.

For networks with under 10,000 nodes set a 30-second timeout.

For over 10,000 nodes or if you have frame relay links, set a 60-second timeout

Never exceed 120 seconds, as you get back to the case of an infinite timeout (try holding your breath for 120 seconds and you’ll get the idea.)

The default timeout for Argent is 30 seconds.

You can carefully adjust the registry entries.

See Also: Argent TCP/IP Timeout Registry Settings

Maximum Allowed Concurrent Connections For TCP/IP Server

Argent engines are implemented as TCP/IP servers.

For large installation, there may be hundreds of clients contacting the server simultaneously. In the case of slow and shared links, you need to control the maximum number of concurrent connections.

Any connection request over the limit is dropped by server immediately.

The default setting is 0, which means there is no limit of concurrent connections.

You can adjust this by carefully editing registries, under registry hive HKEY_LOCAL_MACHINE\Software\Argent

Argent Console

ArgentManagementConsole\TCP_MAX_ALLOWED_CONNECTION

Argent Guardian

ArgentGuardian\TCP_MAX_ALLOWED_CONNECTION

Force Push-out Interval For Argent Console Engine

To prevent stale downloaded data, Argent generates new copy of the BKU file of the shared database periodically. This causes the remote clients to download the new copy of the shared database.

It’s just a typical Argent belt-and-suspenders design – not essential just a nice-to-have for the roughness of the real world.

The default is 30 minutes.

You can adjusted the registry entry:

HKEY_LOCAL_MACHINE\Software\Argent\ArgentManagementConsole\ TCP_FORCE_PUSHOUT_SHARED_DATA

If you set it to zero, then a new copy of the BKU file is created only if the database is modified by user.

Poll Interval For Checking Central Engine For Argent Guardian

The Argent Guardian regional scheduling engine checks the central engine periodically to see if there is a new copy of control information to download.

Based on the frequency of changing control information and the latency you want, you can adjusted the registry

HKEY_LOCAL_MACHINE\Software\Argent\ArgentGuardian\ POLL_CENTRAL_ENGINE_INTERVAL

The default value is 60 seconds.