[I2nsf] Service Level Policies - Post 4: differentiating between context, conditions, and Constraints

[John Strassner <[e-mail]>]

During IETF94, I expressed discomfort with the
"Subject-Object-Action-Function" paradigm. This note defines and
differentiates between the concepts of context, constraints, and conditions.

1.  Context

There are many definitions of context. One of the most popular [1] is:

   ‘‘Context is any information that can be used to characterize the
situation
      of an entity. An entity is a person, place, or object that is
considered
      relevant to the interaction between a user and an application,
including
      the user and application themselves’’.

As stated in [2], the above definition has a number of problems. Instead,
[2] proposes the following definition:

   "The Context of an Entity is a collection of measured and/or inferred
     knowledge that describe the state and the environment in which an
     Entity exists or has existed."

The above definition enables human and or machine inference to be used to
more fully characterize the context of an Entity.

2.  Conditions

An event-condition-action (ECA) policy rule consists of three Boolean
clauses, and has the following conceptual behavior:

   IF the event_clause evaluates to TRUE
      IF the condition_clause evaluates to TRUE
         THEN execute the actions in the action_clause
      ENDIF
   ENDIF

Each of the above clauses may consist of one or more statements. For
example, a condition clause could be:

   if userLogin = Error AND numLoginTries > 3

3.  Constraints

A constraint is a limitation or restriction. Constraints have been used in
programming and in modeling for a long time. Constraint programming refers
to the embedding of constraints in a language; for example, most Prolog
languages include good support for constraint programming. OCL (the Object
Constraint Language) is used to specify constraints in UML.

4. Putting Context, Conditions, and Constraints Together

Given a condition, a constraint can be used to further restrict how to
satisfy the condition. Then, if a context is specified, the condition and
its constraint are bound to that context. In an ECA policy rule, the effect
of this is typically to more fully specify the condition clause.

Note, however, that constraints do not have to be used solely with
Conditions. For example, a constraint can be used to specify when it is
safe to transition to a new state. As another, more powerful, example, the
concept of a software contract [3] uses constraints to extend the
definition of Abstract Data Types to formally define the behavior of
software components. Simplifying the theory, this says the following:

  In order for a method to execute,
      a set of pre-conditions must be satisfied before the method can
execute
      a set of post-conditions must be satisfied when the method is finished
      a set of invariants must not change their value during the execution
-

The set of pre- and post-conditions, along with invariants, are all
constraints.
5. Proposal

Context, conditions, and constraints are separate concepts, but they fit
together naturally.

With respect to policies, if a policy is viewed as a (software) container,
then context is applied to the container (and hence, to the components of
the container).

Constraints can be applied to terms in each of the event, condition, and
action clauses of an ECA policy rule to restrict their behavior, data type,
domain and range, and so forth. The most common examples are applying
constraints to conditions or to govern the behavior of actions.
-

[1] Dey, A., "Providing Architectural Support for Building Context-Aware
     Applications, Ph.D. Thesis, 2000

[2] Strassner, J., et al., "The Use of Context-Aware Policies and
     Ontologies to Facilitate Business-Aware Network Management",
     Journal of Network and Systems Management (17), pg 255-284, 2009
-

[3] Meyer, B., "Object-Oriented Software Construction, 2nd edition,
     Prentice-Hall, ISBN:  0-13-629155-4


-- 
regards,
John