I have trouble understanding the differences between these models.
My understanding:
HRU and Graham-Denning are both computer security models that define the interaction between subjects, objects and actions.
They are similar in that they all define a set of subjects, a set of objects and a an access control matrix.
They are different in regards to how to modify, and who can modify these rights.
It is said that the HRU model is an extension of Graham-Denning model.
Harrison-Ruzzo-Ullman extended this model by defining a system of protection based on commands made of primitive operations and conditions.
And HRU model is more flexible, and is able to describe several access control approaches.
While Graham-Denning model is said to be more limited, with a few problems.
HRU model has six primitive operations:
- grant right to subject and object pair
- delete right from subject and object pair
- create subject
- create object
- delete subject
- delete object Which are then form transitions: sequences of basic operations.
While Graham-Denning defines eight basic rules:
- create an object.
- create a subject.
- delete an object.
- delete a subject.
- grant right.
- delete right.
- transfer right.
- read right.
First Question: What does this difference provide? Does it give more flexibility? (For example, transfer right in G-D can be a transision of deleting and then granting of a right in HRU)
I also know that in G-D each subject has an owner, which can be itself.
Second question: Does the same hold true in HRU?
And when a subject is the owner of itself, then it can delete other subjects access to itself, which is considered undesirable.
Third question: Does HRU solve this problem? (Is it a problem?)
Am I wrong and what can be added?
