DNP3 Driver Communication Statistics

[sbeadle1]

📖 Home  Back  
DNP3 networks (and telemetry systems in general) are prone to unreliable networks where the nature of the protocol allows for a number of retries within each individual message before that message is deemed to have failed.
In the case of large messages, the data returned from the outstation to ClearSCADA may span multiple fragments, where each fragment consists of up to 2048 bytes maximum. Each fragment will be broken up into a number of frames (often 9 or 10) depending on the maximum frame size specified in the ClearSCADA server configuration.

When ClearSCADA requests the data from the outstation, it performs an integrity poll, requesting different classes of data from the RTU. With large amounts of data being returned across sometimes unreliable communications links, it is to be expected that there may be some issues in transferring that data. ClearSCADA has a mechanism where individual frames can be resent (link layer retries).

From an end user perspective, the point of interest is not so much in every frame transmission that may fail, as in the fact that the whole message, after timeouts and retries succeeded or failed. This is mainly due to the unreliable networks possibly causing multiple corrupted frames. These frames can be retried though, meaning that a record of bad messages and/or retried frames does not indicate the real performance of the network.

For DNP3, the points of interest to measure are related to the Application level messages (which includes complete fragments and frames) such as integrity and class polls, and unsolicited messages.

The ClearSCADA database schema details the property names for these statistics. Note that errors can be determined by subtracting the successes from the attempts. These statistics can be found in either the View dialog of the outstation, or from directly from the properties of the outstation which can be placed on a mimic, a trend or report etc.

Note that event polls are not used all that regularly - if they are used the statistics can be retrieved from the properties below.

A second set of statistics are available which document the previous counters stored in the database before the most recent counter reset. Note that these counters return the count between PrevStatsPollStartTime and CurrStatsPollStartTime.

Below is an explanation of the various statistics available on the DNP3 Outstation and Channel objects.

DNP3 Outstation

Statistic	Description
Good Messages	Number of valid data link messages received from the outstation that have a positive response.
No Replies	Number responses that have been timed out, i.e. we have transmitted a message but not received a response (if one was expected) in the time allowed. If the message was only half received then it would count as a no reply & be purged.
Bad Messages	Number of valid data link messages received but whose content is either not correct or contains a negative response. Examples: if the outstation sent a message that had the correct headers and CRCs but whose content did not match what ClearSCADA expected for a reply if the outstation sent a negative response to a control
Errors	Number of data link messages received with invalid CRCs
Integrity Polls Attempted	Number of integrity (Class 0,1,2,3) poll requests sent to this outstation
Integrity Polls Successful	Number of responses received to integrity poll requests from this outstation.
Event Polls (Class n) Attempted	Number of event poll requests sent for a specific event class (Class 1,2 or 3) sent to this outstation.
Event Polls (Class n) Successful	Number of responses received to event poll requests for a specific event class (Class 1,2 or 3) from this outstation
Unsolicited Messages Received	Number of unsolicited responses received from this outstation.

DNP3 Channel

Statistic	Description
Messages	Total number of link layer messages transmitted + valid link layer messages received. This includes the number of broadcast messages transmitted + messages targeted at outstations + good outstation messages received + bad outstation messages received + outstation replies + aborts.
Replies	Total number of messages received as replies to solicited messages (which will include negative responses).
Timeouts	Total number of transmit timeouts (most likely to happen if hardware flow control is enabled) + total outstation no replies.
Errors	Total number of outstation errors.

Note that many of the parameters listed above apply equally to other types of advanced outstation (apart from DNP3 specific Integrity and Event Poll statistics).

Go: Home Back