Download - Academy Publisher

focused mainly on rule-based forms of expression. The
following are examples of some well-known policy
language:
1) PDL (Policy Description Language) is a Bell
Labs-developed event-based language [6]. They use
"event - condition - behavior" rules of the form to define
a policy. PDL’s syntax is simple, with clear semantics
and can express a strong sexual, However, PDL does not
support Access Control Policy, and does not support the
roles of complex strategies.
2) Ponder is a declarative, object-oriented language
for specifying security policies with role-based access
control, as well as general-purpose management policies
for specifying what actions are carried out when specific
events occur within the system or what resources to
allocate under specific conditions [7]. Ponder have four
basic policy types: authorizations, obligations, refrains
and delegations and three composite policy types: roles,
relationships and management structures that are used to
compose policies. The dependencies between the
various types are shown in figure 2. Ponder also has a
number of supporting abstractions that are used to
define policies: domains for hierarchically grouping
managed objects, events for triggering obligation
policies, and constraints for controlling the enforcement
of policies at runtime.
Figure 2.
Policy Types
3) Interval Temporal Logic (ITL) is a linear-time
temporal logic with a discrete model of time. A system is
modelled by a set of relevant state variables An interval
is considered to be a (in)finite, nonempty sequence of
state ∂
0
, ∂
1
,......, where a state ∂
i
is mapping from a
set of variables to a set of values [5]. The first work
investigates logic-based languages for the specification
of authorization is Woo and Lam [8]. Jajodia et al.
worked on a proposal that provides specific rules for
resolving conflicts among positive authorizations and
negative authorizations. Their policy language, called
Authorization Specification Language (ASL), provides a
specific predicate, cando, for expressing explicit
authorizations and a different predicate, dercando, for
expressing derived authorizations [9]. Bertino et al. propose a
model supporting the specification of periodic
authorizations, that is, authorizations that hold in specific
periodic intervals [10]. The syntax of ITL is defined in
Table 1 where:
Expressions
Formulate
TABLE Ⅰ. SYNTAX OF ITL
Z is an integer value,
a is a static integer variable (doesn’t change within an
interval),
A is a state integer variable (can change within an
interval),
v a static or state integer variable,
g is a integer function symbol,
4) XACML is used OASIS (Organization for the
Advancement of Structured Information Standards) to
represent the policy. The current syntax for policy in
XACL is oriented around the 3-tuple {subject, object,
action} [11]. The subject primitive allows user IDs,
groups, and/or role names. The object primitive allows
granularity as fine as a single element within an XML
document. The action primitive consists of four kinds of
actions: read; write; create; and delete. The main
advantage of using XML is: XML is a widely used
standard.
B. Policy language Evaluation Criteria
Each policy language has its own advantages and
disadvantages. Formal Logic language is not enough
intuitive, not easy to map to the realization of the
mechanism, so is difficult to use and understand; Ponder
language is more in line with common standards, could
applies to a variety of applications, but Ponder Toolkit
does not open source; XACML is ideal for on-line
transmission and browse, but have a lot of redundant
information. So when we use policy language, we should
compare their advantages and disadvantages.
Through analysis and research of the existing policy
language, it is obvious that a good policy language
should meet the following requirements:
1) Be easy to understand, The policy written in the
policy language should be easy to understand, which is
very important. Only when the policy could easily
understood can achieve a wide range of applications.
2) Avoid absolutes, It is not easy to map abstract
representations into the implementation mechanism, thus
too abstract representations should be avoided.
3) Be do-able, It is easy to map to a description of the
underlying measurement specifications, it is compiled
into some kind of low-level (e.g. C) language code.
4) Good scalability, According to demand, it can mix
new policy types with the existing policy , namely,
support for complex strategies, and supports reusability。
5) Policy classification high covering,The language
can cover all the current international standards of
information security policy.
215