1. 12/13/20151 Computer Security Security Policies..

2. 12/13/20152 Security Policies We view a computer system as a finite-state machine Definition A security policy is a statement that partitions the states of a system into a set of authorized or secure states and a set of unauthorized or nonsecure states. A secure system is a system that starts in an authorized state and cannot enter an unauthorized state.

3. 12/13/20153 Example s 1 s2s2 s3s3 s4s4 t1t1 t3t3 t2t2 t5t5 t4t4 An insecure system Authorized states are s 1 and s 2 Unauthorized states are s 3 and s 4

4. 12/13/20154 Security Policies Definition 1.A breach of security occurs when a system enters an unauthorized state. 2.Let X be a set of entities and I be some information. I has the property of confidentiality with respect to X if no member of X can obtain information about I. I has the property of integrity with respect to X if all members of X trust I. 3.Let I be a resource I has the property of availability with respect to X if all member of X can access I. 5.A security mechanism is an entity or procedure that enforces some part of a security policy.

5. 12/13/20155 Types of Policies Definition 1.Military security policies or governmental security policies. 2.Commercial security policies Confidentiality policies Integrity policies Transaction policies Discuss issues regarding trust.

6. 12/13/20156 The role of trust The role of trust is fundamental in understanding the nature of computer security. Examples –see textbook [Example 1-2-3-4, pp 101-102 (high level)& 1-2-3-4 102-103 (low-level, formal)]

7. 12/13/20157 Types of Access Control Discretionary Access Control (DAC) or identity based access control. Mandatory Access Control (MAC) or role-based access control. An originator access control (ORCON or ORGON) bases access on the creator of an object. [Examples pp 103-104]

8. 12/13/20158 Discretionary Access Control (DAC) Access control is left to the discretion of the owner. Based on the identity of the subject. [Example –see textbook pp 104-105

9. 12/13/20159 Mandatory Access Control (MAC) The operating system enforces mandatory access controls. Neither the subject nor even the owner can determine access control. Example –see textbook

10. 12/13/201510 ORiginator access CONtrol (ORCON or ORGON) The originator of the file (or its information) has control the dissemination of its information. Example –see textbook

11. 12/13/201511 Policy languages High level policy languages: independent of the mechanisms used. Low level policy languages [Examples pp 104-105]

12. 12/13/201512 High level policy languages Express policy constraints on entities using abstraction and are independent of the security mechanisms. This requires: An unambiguous expression of policy A mathematical or programming formulation Details: see textbook. [Examples pp105-106]

13. 12/13/201513 Low level policy languages A set of inputs or arguments to commands that set or check constraints on a system. For examples, see textbook [Examples pp109-110]

14. 12/13/201514 Security and Precision Earlier security and precision was defined in terms of the states of the system. We said that security policies were enforced by security mechanisms and that such mechanisms were either secure, precise or broad. Let P be the set of all states, Q the set of secure states and suppose that the mechanism restricts the system to the set of states R. A security mechanism was secure if R  Q, precise if R = Q and broad if there are states such that r  R and r  Q.

15. 12/13/201515 Security and Precision We now consider the possibility of devising a generic procedure for developing a mechanism that is security and precise. For this, we will use programs, which will be viewed as abstract functions that “encode” the information that needs to be controlled.

16. 12/13/201516 Security and Precision Definition A program p is a function p : I 1  …  I n → R. p has n inputs i j  I j and one output r  R Axiom ( observability postulate ) Suppose p does not alter information but merely provides a view of its inputs. We say that p encodes all available information about i 1, …, i n Example A confidentiality policy seeks to control what views are available.

17. 12/13/201517 Security and Precision Definition Let p : I 1    I n → R be a function 1. A protection mechanism m for p is a function m : I 1   I n → R  E ( E is an error message ) for which, when ( i 1  i n )  I 1   I n, either a. m ( i 1  i k  = p ( i 1  i k  or b. m ( i 1  i k   E. That is, every “legal” input to m produces either the same value as p or an error message. The set of output values of p that are excluded by m are those outputs that would impart confidential information. [Examples p 115]

18. 12/13/201518 Security and Precision Definition 2. A confidentiality policy for the program p : I 1  I n → R is a function c : I 1   I n → A, where A is a subset of I 1   I n. Here the set A corresponds to those inputs that may be revealed. The complement of A to the confidential inputs.

19. 12/13/201519 Security and Precision Definitions 3.Let c be a confidentiality policy for a program p. Let m : I 1   I n → R  E be a security mechanism for p. The mechanism m is secure iff there is a function m’ : I 1   I n → R  E such that for all ( i 1  i n )  I 1   I n   m ( i 1  i k  = m’ ( c ( i 1  i k )). That is, given any set of inputs, the protection mechanism m returns values consistent with the stated policy c ( here “secure” = “confidential” )

20. 12/13/201520 Security and Precision Definitions Let m 1, m 2 protection mechanisms for program p under policy c. 4. m 1 is as precise as m 2 if for all inputs ( i 1  i n ) : m 2 ( i 1  i k  = p ( i 1  i k  ═> m 1 ( i 1  i k  = p ( i 1  i k  5.m 1 is more precise than m 2 if there is an input ( i 1 ′  i n ′ ) such that : m 2 ( i 1 ′  i n ′ ) = p ( i 1 ′  i n ′ ) & m 1 ( i 1 ′  i n ′ ) ≠ p ( i 1 ′  i n ′ )

21. 12/13/201521 Security and Precision Theorems 1.For any program p there exists a precise secure mechanism m* such that for all secure mechanisms m associated with p and c we have m* = m. 2.There is no effective way that determines a (maximally) precise secure mechanism for any policy and program.