RE: [bt-spec] BTP Issue 21 : Multiple Superior:Inferior relationshipswithin a Cohesion

[Peter Furniss <peter.furniss@choreology.com>]

 

BTP Issue 21 : Multiple Superior:Inferior relationships within a Cohesion

Category: minor technical
Submitter: Gordon Hamilton, AppliedTheory
Description:
Is it an issue that an inferior may be in multiple S:I relationships within a BT Cohesion and cancel or hazard one of them while confirming another? Or is this just an implementation detail for the Terminator functionality, where they can choose to monitor at that granular a level?

---

 

? 

 

An Inferior, as *a BTP role*, is a one transaction thing, involved with a single Superior for the lifetime of one business transaction.  (In the summer, there was considerable discussion on the question of whether there is one or several Participants in a host, which is the same point; as a BTP role, "Participant" means the one-transaction thing, even if the software process and the address (but not the identifier) are involved in multiple transactions, concurrently or consecutively)

 

So the software process (or whatever), may indeed be enrolled multiple times, but it is then viewed as multiple Inferiors. These may be in the same Cohesion (or the same Atom - just because it wasn't necessary to enrol another Inferior doesn't mean it is disallowed)

 

If the business transaction (Superior) is a Cohesion, then have to be separate Inferiors, and they will be liable to get different results. The Terminator in this case will almost certainly be "monitoring at that level". Typically, one would expect the Terminator to have some understanding (via application means, assisted perhaps by the "infeiror handle" qualifier) of what application work was under the control of which logical Inferior.  (consider the choreology demo - the stock and option from the same market maker are separate Inferiror's hosted in the same server. The Terminator can "see" the two Inferiors and "knows" which is which - so if the user chooses stock from A and option from B, the right Inferiors are included in the confirm set.

 

I suggest no change to the text.