Robert,
Congratulations on exploiting the extended value of the SMAT files. I personally think there is a lot of valuable information in those files to support various analyses.
Here is my input/opinion:
The MID is the same Message ID that is in the Message METADATA within the engine at the time the SMAT message is written. I believe that id can run to a billion and so may eventually roll over but should remain relatively unique unless there is some way to reset the counter used to create MIDs or you have an exceptional volume of messages.
I think that number is also related to the number of Sites you have (I think it is related to the Recovery DB and you get one of those per site). So if I am correct, it s quite possible that with multiple sites, SMAT files from different sites could have the same MID for different messages. So you might want to include in your extracted file some indication of the site.
Actually that should not have been a problem but Cloverleaf(R) has not yet utilized the full MID number (I think it is domain.hub.mid). The domain was intended to indicate a platform (say you had multiple machines running Cloverleaf(R) each would get its own domain id); hub was suposed to indicate the Site; and then finally the mid for the mesage within that site within that machine.
Unfortunately the domaind and hub indicators are always zero (so we see 0.0.123456789.
At one time there was talk of having multiple platform, multiple site support naturally deployed in Cloverleaf(R) (and this MID construct is but one indication the infrastructure is in place) so that we would not have to actually define intersite connections for example but that has not yet occurred.
The natural multi-site multi-domain support within Cloverleaf(R) is something I requested nearly 10 years ago (when Oakwood, where I then worked) was among the first customers to deploy multiple site architecture.
The TYPE is probably DATA or REPLY and again is what is in the same METADATA field of the message at the time the SMAT file is written.
The DATAFMT is not reliably completed. I do not recall if there is a METADATA equivalent. Ideally it should indicate the format (such as variant name, frl name, vrl name, etc.) that is associated with the message within the engine – and sometimes it does. I have asked about making that reliable and it was indicated that that could not always be determined. Frankly, I find that hard to believe.
I do not know that there is any documentation regarding these data elements other than when there is a METADATA element being repeated in the Index file then the METADATA documentation would probably apply.
I think to do a thorough analysis one would almost have to also use the MSI stats as well as SMAT based stats (each perhaps for different purposes).
Jim Kosloskey
email: jim.kosloskey@jim-kosloskey.com 29+ years Cloverleaf, 59 years IT - old fart.
.
