Clarification on KB article about DNP3 AL timeouts between Geo SCADA and outstation

[rlao]

This KB article outlines how DNP3 comms issues can be caused if the AL timeouts between Geo SCADA and the outstation are not in sync.

A snippet of the article states that:

"The issue arises when ClearSCADA's AL timeout is set to be equal to or smaller than that in the outstation. The (old) DNP3 rules stated that an outstation will not respond to a new Read request if it has a pending response that requires a confirm."

However, the conclusion then states:

"To prevent this issue from occurring, the outstation AL timeout should be greater than the ClearSCADA AL timeout (usually by at least a couple of seconds)."

My questions are:

• Shouldn't the outstation AL timeout be shorter than the Geo SCADA AL timeout by a couple of seconds, rather than longer? My understanding is that we want the outstation to time out first.
• In a SCADAPack outstation object, what is the relevant field that this applies to? Is it the Master 1 configuration 'Application Layer Confirm (Event) Timeout' or the Application Layer 'Complete Fragment Timeout' or both that should be synced to the Geo SCADA AL timeout?
• By a couple of seconds, will 1 second suffice? Should it be more? Or does it depend on the system?