I don't know whether its correct or not.
As per my interpretation, it means all schedules possible under protocol of particular oval is also possible under protocol of enclosing oval.
However as long as I know, both multiversion timestamp and timestamp ordering protocols ensures conflict serializability, whereas Thomas write rule ensures view serializability.
So I guess, the diagram correctly depicts the relation between Thomas write rule and timestamp ordering protocol. But I am unsure whether it correctly depicts the relation of multiversion timestamp ordering protocol correctly in relation with Thomas write rule. (I believe relation between multiverstion timestamp ordering protocol and timestamp ordering protocol is also correctly shown).
Should it be something like this:
I have came with above diagram because I feel it correctly captures following points:
- All schedules possible under timestamp ordering are also possible under multiversion timestamp
- All schedules possible under timesstamp ordering are also possible under thomas write rule
- Some schedules (specifically view serializable ones) possible under Thomas write rule are not possible under multi version timestamp protocol and vice versa (schedules involving different versions of data items).
Am I right with all of these?