Lecture 10 - The University of Texas at Dallas

Download Report

Transcript Lecture 10 - The University of Texas at Dallas

Cyber Security
Lecture for June 25, 2010
Unit #2: Selected Topics in
Cyber Security
Dr. Bhavani Thuraisingham
The University of Texas at Dallas
June 25, 2010
Outline
 Operating Systems Security
 Network Security
 Designing and Evaluating Systems
 Web Security
 Data Mining for Introduction Detection
 Other Security Technologies
Operating System Security
 Access Control
- Subjects are Processes and Objects are Files
- Subjects have Read/Write Access to Objects
- E.g., Process P1 has read acces to File F1 and write access to
File F2
 Capabilities
- Processes must presses certain Capabilities / Certificates to
access certain files to execute certain programs
- E.g., Process P1 must have capability C to read file F
Mandatory Security
 Bell and La Padula Security Policy
- Subjects have clearance levels, Objects have sensitivity levels;
clearance and sensitivity levels are also called security levels
- Unclassified < Confidential < Secret < TopSecret
- Compartments are also possible
- Compartments and Security levels form a partially ordered
lattice
 Security Properties
- Simple Security Property: Subject has READ access to an object
of the subject’s security level dominates that of the objects
- Star (*) Property: Subject has WRITE access to an object if the
subject’s security level is dominated by that of the objects\
Covert Channel Example
 Trojan horse at a higher level covertly passes data to a Trojan
horse at a lower level
 Example:
- File Lock/Unlock problem
- Processes at Secret and Unclassified levels collude with
one another
- When the Secret process lock a file and the Unclassified
process finds the file locked, a 1 bit is passed covertly
- When the Secret process unlocks the file and the
Unclassified process finds it unlocked, a 1 bit is passed
covertly
- Over time the bits could contain sensitive data
Steps to Designing a Secure System
 Requirements, Informal Policy and model
 Formal security policy and model
 Security architecture
- Identify security critical components; these components must be
trusted
 Design of the system
 Verification and Validation
Product Evaluation
 Orange Book
- Trusted Computer Systems Evaluation Criteria
 Classes C1, C2, B1, B2, B3, A1 and beyond
- C1 is the lowest level and A1 the highest level of assurance
- Formal methods are needed for A1 systems
 Interpretations of the Orange book for Networks (Trusted Network
Interpretation) and Databases (Trusted Database Interpretation)
 Several companion documents
- Auditing, Inference and Aggregation, etc.
 Many products are now evaluated using the federal Criteria
Network Security
 Security across all network layers
- E.g., Data Link, Transport, Session, Presentation,
Application
 Network protocol security
Ver5ification and validation of network protocols
 Intrusion detection and prevention
- Applying data mining techniques
 Encryption and Cryptography
 Access control and trust policies
 Other Measures
- Prevention from denial of service, Secure routing, - - -
-
Data Security: Access Control
 Access Control policies were developed initially for file systems
- E.g., Read/write policies for files
 Access control in databases started with the work in System R and
Ingres Projects
- Access Control rules were defined for databases, relations,
tuples, attributes and elements
- SQL and QUEL languages were extended

GRANT and REVOKE Statements

Read access on EMP to User group A Where
EMP.Salary < 30K and EMP.Dept <> Security
- Query Modification:

Modify the query according to the access control rules

Retrieve all employee information where salary < 30K and
Dept is not Security
What is an MLS/DBMS?
 Users are cleared at different security levels
 Data in the database is assigned different sensitivity levels--
multilevel database
 Users share the multilevel database
 MLS/DBMS is the software that ensures that users only
obtain information at or below their level
 In general, a user reads at or below his level and writes at his
level
Why MLS/DBMS?
 Operating systems control access to files; coarser grain of
granularity
 Database stores relationships between data
 Content, Context, and Dynamic access control
 Traditional operating systems access control to files is not
sufficient
 Need multilevel access control for DBMSs
Summary of Developments in MLS/DBMS
 Early Efforts 1975 – 1982; example: Hinke-Shafer approach
 Air Force Summer Study, 1982
 Research Prototypes (Integrity Lock, SeaView, LDV, etc.);
1984 - Present
 Trusted Database Interpretation; published 1991
 Commercial Products; 1988 - Present
Inference Problem
 Inference is the process of forming conclusions from premises
 If the conclusions are unauthorized, it becomes a problem
 Inference problem in a multilevel environment
 Aggregation problem is a special case of the inference
problem - collections of data elements is Secret but the
individual elements are Unclassified
 Association problem: attributes A and B taken together is
Secret - individually they are Unclassified
Security Threats to Web/E-commerce
Security
Threats and
Violations
Access
Control
Violations
Denial of
Service/
Infrastructure
Attacks
Integrity
Violations
Fraud
Sabotage
Confidentiality
Authentication
Nonrepudiation
Violations
Data Mining for Intrusion Detection: Problem

An intrusion can be defined as “any set of actions that attempt to
compromise the integrity, confidentiality, or availability of a resource”.

Attacks are:

Intrusion detection systems are split into two groups:

Host-based attacks
Network-based attacks
Anomaly detection systems
Misuse detection systems
Use audit logs
-
Capture all activities in network and hosts.
But the amount of data is huge!
Misuse Detection
 Misuse Detection
Problem: Anomaly Detection
 Anomaly Detection
Other Security Technologies
 Digital Identity Management
 Identity Theft Management
 Digital Forensics
 Digital Watermarking
 Risk Analysis
 Economic Analysis
 Secure Electronic Voting Machines
 Biometrics
 Other Applications
Digital Identity Management
 Digital identity is the identity that a user has to access an
electronic resource
 A person could have multiple identities
- A physician could have an identity to access medical
resources and another to access his bank accounts
 Digital identity management is about managing the multiple
identities
- Manage databases that store and retrieve identities
- Resolve conflicts and heterogeneity
- Make associations
- Provide security
 Ontology management for identity management is an
emerging research area
Digital Identity Management - II
 Federated Identity Management
- Corporations work with each other across organizational
boundaries with the concept of federated identity
- Each corporation has its own identity and may belong to
multiple federations
Individual identity management within an organization
and federated identity management across organizations
 Technologies for identity management
- Database management, data mining, ontology
management, federated computing
-
Identity Theft Management
 Need for secure identity management
- Ease the burden of managing numerous identities
- Prevent misuse of identity: preventing identity theft
 Identity theft is stealing another person’s digital identity
 Techniques for preventing identity thefts include
- Access control, Encryption, Digital Signatures
- A merchant encrypts the data and signs with the public
-
key of the recipient
Recipient decrypts with his private key
Digital Forensics
 Digital forensics is about the investigation of Cyber crime
 Follows the procedures established for Forensic medicine
 The steps include the following:
- When a computer crime occurs, law enforcement officials
-
who are cyber crime experts gather every piece of
evidence including information from the crime scene (i.e.
from the computer)
Gather profiles of terrorists
Use history information
Carry pout analysis
Digital Forensics - II
 Digital Forensics Techniques
- Intrusion detection
- Data Mining
- Analyzing log files
- Use criminal profiling and develop a psychological
profiling
- Analyze email messages
 Lawyers, Psychologists, Sociologists, Crime investigators
and Technologists have to worm together
 International Journal of Digital Evidence is a useful source
Steganography and Digital Watermarking
 Steganography is about hiding information within other
information
- E.g., hidden information is the message that terrorist may
be sending to their pees in different parts of the worlds
- Information may be hidden in valid texts, images, films
etc.
- Difficult to be detected by the unsuspecting human
 Steganalysis is about developing techniques that can analyze
text, images, video and detect hidden messages
- May use data mining techniques to detect hidden patters
 Steganograophy makes the task of the Cyber crime expert
difficult as he/she ahs to analyze for hidden information
- Communication protocols are being developed
Steganography and Digital Watermarking - II
 Digital water marking is about inserting information without
being detected for valid purposes
- It has applications in copyright protection
- A manufacturer may use digital watermarking to copyright
a particular music or video without being noticed
- When music is copies and copyright is violated, one can
detect two the real owner is by examining the copyright
embedded in the music or video
Risk Analysis
 Analyzing risks
- Before installing a secure system or a network one needs
to conduct a risk analysis study
- What are the threats? What are the risks?
 Various types of risk analysis methods
Quantitative approach: Events are ranked in the order of
risks and decisions are made based on then risks
Qualitative approach: estimates are used for risks
-
Economics Analysis
 Security vs Cost
- If risks are high and damage is significant then it may be
worth the cost of incorporating security
- If risks and damage are not high, then security may be an
additional cost burden
 Economists and technologists need to work together
- Develop cost models
- Cost vs. Risk/Threat study
Secure Electronic Voting Machines
 We are slowly migrating to electronic voting machines
 Current electronic machines have many security
vulnerabilities
 A person can log into the system multiple times from different
parts of the country and cast his/her vote
 Insufficient techniques for ensuring that a person can vote
only once
 The systems may be attacked and compromised
 Solutions are being developed
 Johns Hopkins University is one of the leaders in the field of
secure electronic voting machines
Biometrics
 Early Identication and Authentication (I&A) systems, were
based on passwords
 Recently physical characteristics of a person are being sued
for identification
- Fingerprinting
- Facial features
- Iris scans
- Blood circulation
- Facial expressions
 Biometrics techniques will provide access not only to
computers but also to building and homes
 Other Applications
Biometric Technologies
 Pattern recognition
 Machine learning
 Statistical reasoning
 Multimedia/Image processing and management
 Managing biometric databases
 Information retrieval
 Pattern matching
 Searching
 Ontology management
 Data mining
Data Mining for Biometrics
 Determine the data to be analyzed
- Data may be stored in biometric databases
- Data may be text, images, video, etc.
 Data may be grouped using classification techniques
 As new data arrives determine the group this data belongs to
- Pattern matching, Classification
 Determine what the new data is depending on the prior
examples and experiments
 Determine whether the new data is abnormal or normal
behavior
 Challenge: False positives, False negatives
Secure Biometrics
 Biometrics systems have to be secure
 Need to study the attacks for biometrics systems
 Facial features may be modified:
- E.g., One can access by inserting another person’s
features
Attacks on biometric databases is a major concern
 Challenge is to develop a secure biometric systems
-
Secure Biometrics - II
 Security policy for as biometric system
- Application specific and applicatyion independent
policies
- Security constraints
 Security model for a biometrics systems
Determine the operations to be performed
- Need to include both text, images and video/animation
 Architecure foe a biometric system
- Need to idenify securiy critical components
Reference monitor
 Detecting intrusions in a biometric system
-
-
Other Applications
 Email security
- Encryption
- Filtering
- Data mining
 Benchmarking
- Benchmarks for secure queries and transactions
 Simulation and performance studies
 Security for machine translation and text summarization
 Covert channel analysis
 Robotics security
- Need to ensure policies are enforced correctly when
operating robots