In part 1 I discussed the new Windows Workflow features in Metastorm BPM 7.6. I'll now move onto the database tables used by WF workflows in Metastorm.
First we have two tables that contain process metadata.
eMSWorkflow is pretty simple, it contains one entry for each workflow published to the database. There is a column for the name of the workflow and a GUID column. The name is the name that will be used when executing a workflow from Metastorm. I'm unclear what the GUID column is for. Multiple workflows with the same name aren't possible so it isn't for differentiating between two workflows with the same name.
eMSWorkflowDefinition holds an entry for each workflow version published to the database. Most of this table is pretty self-explanatory. Again there is a GUID column whose purpose isn't clear. There is an interesting column called 'eWorkflowDefinitionType'. I haven't investigated this too far but it suggests that it is possible to publish workflows as assemblies or as uncompiled XOML file.
Next we have tables that hold workflow instance data.
eMSWorkflowTracking contains audit trail information for executing workflows. It looks like Metastorm have implemented their own tracking service which makes sense since it means the workflow audit trail can be connected to the Metastorm folder's audit trail. One problem I have with the current implementation is there appears to be no way to connect a workflow instance to its type, so it will be difficult to figure out how many instances of a particular workflow type are in existence.
eMSWorkflowEvent also includes data about workflow instances. The difference is that this table contains information that is related to the workflow runtime, rather than the workflow itself. So it will contain entries for when an instance has been loaded or idled etc.
Finally there are two other tables. These are the standard tables created for the WF persistence service, CompletedScope and InstanceState. The Microsoft documentation for these tables seems to be pretty much non-existent, so I don't have much idea what is in there. All I can assume is the 'state' column contains the serialized data of the current state of the workflow instance.
So that's the database tables. Part 3 will discuss running workflows generated in Visual Studio (if I get time to actually get it working).