[tm-rg] GGF14 Meeting and Group Sunset
Haugen Robert
Robert.Haugen at choreology.com
Thu Jun 23 09:32:52 CDT 2005
Hi Torsten,
Sorry I can't attend the phone meeting tomorrow.
Comments on slides:
Typo on slide 5:
"Distributed avaiable-to-promise" should be "Distributed
available-to-promise"
Slide 15:
"Distributed 2PC
Scaleability Problem in terms of resource sharing"
I think that statement incorrectly equates 2PC with ACID rules and
holding resource locks for the duration of a transaction.
Several of the loosely-coupled transaction protocols allow distributed
transactions which have two phases but which do not require holding
locks and relax some ACID rules.
Compensation is actually a 2-phase outcome, following the do-compensate
pattern.
Anything following the provisional-final pattern has 2 phases but does
not require holding locks.
Options are an implementation of the provisional-final pattern.
More information about the tm-rg
mailing list