Smeal PhD Renewal Report - Information Systems and

Download Report

Transcript Smeal PhD Renewal Report - Information Systems and

Information System Security
Engineering and Management
Dr. William Hery
[email protected]
[email protected]
CS 996
Spring 2005
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
Outline of Presentation
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
• Course Motivation
• Approach to Learning, Grading in This
Course
• Main Course Topics
• Highlights of course topics to show
linkage
• Term Project Structure
Initial Course Motivation
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
• For SFS students: fill in gaps in National Security
Telecommunications and Information Systems Security
Committee (NSTISSC) certification for NSA
 NSTISSI 4011: National Training Standards for INFOSEC
Professionals (http://www.nstissc.gov/html/library.html)
• Most technical topics are covered in other courses
 Missing NSTISSI technical tidbits inserted as needed
• The “missing” topics are all related to management, policy and
systems engineering.
 The course will be a survey of information system security engineering
and management topics over a system life cycle
• Although a “government” motivation in selecting the topics,
all are broadly applicable to developing and managing
commercial systems securely
Course Focus
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
• Broad management perspective applicable to DoD/NSA, civilian
government agencies, corporate world: think like a manager
 If you are a manager
 If you have to deal with a manager
• System, not detail, focus
 Not about security products (crypto, fiewall, etc.), but how to use them
in a system
• Many topics are subjective, not objective
 There may be no “right way” or “right answer”
• Many topics should be courses in themselves
 This course will teach you what to think about, not how to do everything!
• Secondary goal: Gain experience in teamwork, government project
organization, presentations and report writing.
Course Organization
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
• Weekly graded homework
• Each student will present a one hour lecture on
a topic--and assign reading and homework for it
• Reading assignments and class discussion
 Active participation in discussion part of grade!
• Student team projects (more later)
Grading
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
 Homework:
 Due before the next class after assignment
 Each graded on a 100 point scale
 10 points per week (or fraction thereof) deducted for any homework
submitted after the due date
 Ten highest individual HW grades averaged to get overall HW grade
 Lecture: Graded on the basis of discovering and understanding
material, and organization of presentation.
 Team Project: Grades based on:
 Depth of understanding of security both the technical issues and the
application of systems engineering and management processes
 Organization of final presentation and report
 All members of a team get the same grade
 Overall grade: 40% homework, 40% project, 20% lecture
References
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
• Primary text: Ronald Krutz and Russell Vines,
The CISM Prep Guide, Wiley, 2003, ISBN 0-47145598-9
• Supplementary material from:
 Ross Anderson, Security Engineering, Wiley,
2001, ISBN 0-471-38922-6
 Tipton and Krause, Information Security
Management Handbook, 4th Edition,
Auerbach, ISBN 0-8493-1518-2 (Copy in ISIS
Lab)
 Various web sites, etc.
Key Points About Security For a
Manager
•
Information Security is a key part of the life cycle processes for any
system:




•
•
•
•
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
System conception and design
System development
Operation of deployed systems
System “decommission”
It is critical to include security considerations from the beginning of system
conception and design
Use of security technology (firewalls, crypto, IDS, patching), etc. is only
part of security: security is based on people, processes, and technology
Actively managing the “security process” is a key part of achieving security
People (developers, users, hackers) are (sometimes without actively
knowing it) are part of the security process
Key Points (Part II)
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
• Security is a wide array of options, not a yes/no choice
 No security is probably not enough
 Near perfect security is difficult, expensive, hard to use, and
takes a long time to do. It is probably too much
• The key is to find the “sweet spot” of enough security
for your particular system. This is a key management
decision.
• Many topics are subjective, not objective
 There may be no “right way” or “right answer”
Extremes of Security
A system in an open
area, which allows
anyone to use it and
allows anything to come
in or go out over multiple
networks provides
virtually no security
(e. g., Internet Café)
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
A system in a locked room, with
2 foot thick concrete and metal
walls (Faraday cage), no
windows, battery operated (no
power lines, a secure operating
system, strongly encrypted
data, armed guards, and only
one person allowed in provides
very strong security (e. g., most
secure NSA system…maybe)
Costs of Security
Security
Development cost
Development time
Equipment cost
Maintenance cost
Usability?
Functionality?
Performance?
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
Finding the “Sweet Spot”
•
•
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
The sweet spot depends on what you need to protect, and what
environment you are operating in
The key to finding the “sweet spot” is to understand the security
requirements
 Requirements may be hard to pin down
 Different aspects of a system may have very different security requirements
 The “sweet spot” is getting the right level of security for each aspect
•
Two common errors in finding the “sweet spot”:
 Uniformly low security because management does not understand the risk.
 Uniformly, strong, and expensive approach to security everywhere, when less
(e. g., no crypto) is enough in most places, with something stronger in selected
places (e. g., protecting passwords or crypto keys)
What is Information Security?
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
• A set of properties of the information system, not a
technology
• These properties are provided with both of processes
and technologies
• The properties: CIA
 Confidentiality: only permitted entities are allowed to “see” the
information
 Integrity: only permitted entities are allowed to modify the
information (this includes creation and deletion)
 Integrity preservation: you know it can’t be changed
 Integrity violation detection: you can’t trust and must go to a
backup or alternate source
 Availability: the information is available when needed
Related security concepts
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
• Identification: a means of saying who/what an entity is
• Authentication: a means to verify that an entity is who it claims to
be for decisions in support of confidentiality and integrity
• Access Control: a means to enforce which entities have access to
information to support confidentiality and integrity
• Authorization: a combination of authentication (who) and access
control
• Non-repudiation: integrity of the pair (information, creator of
information)
• Privacy: confidentiality of personal information
• Anonymity: confidentiality of identity
• Recovery: restoration of a system to a “correct” state after a
security incident.
Methods to Provide Security Properties
S ecu rity P r op er ty
C on fi d en tia li ty
In te gr ity pre se rv at ion
In te gr ity de tec tio n
A va il a bi lit y
S am ple No n-IT M etho d
S ea led en ve lo p e, Sa fe
S af e
“S afet y ” ch ec k b ac k grou nd , Wa term a rk
M ultip le cr e dit c ar d s (av ai lab ili ty o f
p er son al cr e dit)
Id en tific a tio n
A uthe n tic atio n
N am e tag s
S ig n atur e , ID ba dg e, p assp or t
A cc es s C on tro l
N on -R ep ud ia tio n
L oc ks , ID b ad ge s
N otar y P ub li c ; c er tifi e d m a il w ith re tur n
re ce ip t (n on -r e pu diat ion of se nd in g , a nd
re ce ip t)
L oc ke d s af e at hom e, sh re d ding
d oc um en ts
C al l from c oi n p ho ne
P ri v ac y
A no nym it y
R ec ov er y
P ho toc op ie s o f i mp ortan t d oc um en ts in
sa fe de po sit bo x
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
S am ple IT T e ch no lo gy
C ry pto , A cc ess c on tro l
A cc es s C on tro l, w rit e on ly m ed ia
C ry pto , sec ur e h as h
R ed un da nt se rve r l o ca tio ns; stro ng sy stem s
se cu ri ty (fi rew al l, ID S , sec ur e d O S &
ap plic a tio ns, e tc.)
L og in ID , u se r n am e
ID /p assw ord, bi om et ric s (fin g erprin t, r e tin al
sc an , e tc.), c ry pto to ke ns, d ig it a l c er tifi c ate
L og in /pa sswo rd , file p erm issio n s
D ig ita l S ig n atur e
E nc ry p te d fil e s o n h ome com pu ter
A no nym ou s w eb b ro w sing se rv ic e s (e. g.
h ttp :// www .p riv o xy .o rg /)
O ff sit e ba ck up a t s tan db y fac ili ty
DoD terminology
•
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
Communications Security (COMSEC)
 Security of information (voice, data) while in transit. Includes switched circuits,
radio links, microwave, satellite, packet nets, Asynchronous Transfer Mode
(ATM), Synchronous Optical Networks (SONET), Packet over fiber, free space
optics, etc.
•
Computer Security (COMPUSEC)
 Security of information while stored or being processed on a computer
•
Information Security (INFOSEC)
 COMPUSEC + COMSEC
•
Transmission Security (TRANSEC)
 Security of Transmission media
•
Operations Security (OPSEC)
 Operational processes for protecting potentially sensitive unclassified material
(people and technology)
•
Automated Information Systems (AIS)
 Computers + networks linking computers
Security vs. Reliability
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
• Security attacks, software flaws, and hardware
failure can all lead to violations of “CIA”
• For some events, it may be hard to determine
which class of flaws is the cause.
• Some protection and recovery mechanisms are
the same for both security attacks and
hardware or software failures
Security vs Reliability Differences
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
• Hardware failures
 No malicious cause
 Usually affects “A”, sometimes “I” or “C”
 Typically independent events
 Testing is often a reliable way to find hardware failures on deployed
systems
 Stochastic and temporal (e. g., mean time between failure, MTBF)
failure models are useful metrics
 “Availability” is also a standard term in reliability
• Software failure
 No malicious attack: design or coding error
 Can affect “A”, sometimes “I” or “C”
 Often correlated events from same flaw as similar state
conditions arise in different instantiations
 Stochastic models of limited value
Security vs Reliability Differences
(continued)
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
• Security breach
 Malicious attack
 Serious attacks often attempt to hide event
 Can affect “A”, “I” or “C”
 In most cases, the most serious impacts are
attacks on “I” or “C”
 Many attacks are highly correlated
worldwide, but some are very targeted and
correlations may be hard to find
Survivable Systems
•
•
•
•
•
•
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
Systems that provide both reliability and security are called survivable (or
dependable)
Reliability and security requirements are often similar in nature
(particularly availability), and it makes sense to combine the requirements
analysis for both
Designing secure systems and reliable systems both depend on
understanding what is at risk if there is a failure (security compromise or
system failure), what the threats are (hackers, failure modes), and
managing that risk
It is sometimes hard to distinguish between a reliability failure and a
security breach (e. g., the 2003 northeast blackout).
Recovery from a failure and a security breach are sometimes similar, and
it makes sense to combine the recovery plans.
Security is considered by some to be a subset of reliability, with security
breaches just another form of failure…
 But the malicious, planned, correlated, and hidden aspects of security breaches
requires a very different protection approach to most aspects of reliability
Where are security flaws?
•
In system design


•
Ambiguous/incomplete design document
Implementation errors (buffer overflows, etc.)
In system use



•
Not planning for security (many things today)
Designing security incorrectly (WiFi original encryption standard)
In system implementation


•
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
Configuration--often weakest security “out of the box”
Failure to keep up with updates/patches
Physical security
Ill advised user actions


Poor passwords/passwords written down
Victims of “social engineering”
Management needs to keep all of this in mind when designing,
implementing and deploying systems
Management Security Concerns
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
• Classified information at DoD/NSA/other govt agencies:
 National security, loss of life, “sources and methods,” political,
career impacts of security breech
• Unclassified government information:
 Political, financial, legal, career impacts of security breech
• Corporate
 Financial, intellectual property, legal, corporate image, career
impacts of security breech
• Almost no managers: neat technology
Sources of Security Requirements
•
•
•
•
•
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
Risk analysis (national security, lives, property, money)
Legal (e. g., HIPAA, Sarbannes_Oxely, privacy laws)
Higher level government/corporate policies
Corporate/agency/personal image
Others derived from the above
• Requirements may change due to costs, changing
threat environment, etc.
 Requirements may not be known or understood at the start of a
project
Steps to Include Security in
the System Life Cycle
•
•
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
Risk analysis (for the most fundamental security requirements)
Complete security requirements analysis
 Security is a “non-functional” requirement, as is reliability
•
•
High level security policy (technology, management processes, personnel
policies)
Overall system engineering
 Includes security design and development
 Lower level security requirements and policies developed
 Security should be an integral element from the start
•
•
•
•
Security management of deployed system
Incident Response
Business Continuity Planning
Decommissioning of systems and components
Risk Analysis
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
• What is at risk (national security, lives, property, money)?
 Some risk models are based on $ values
• Where does the threat come from?
 Motivation (national security, money, fame,
 Capabilities (intellect, equipment, money)
• What vulnerabilities can be exploited
 Technical
 Process
 People
• Risk management
 Eliminate/reduce risk (e. g., put in crypto, firewall…)
 Accept risk (with recovery process)
 Transfer risk (e. g., to an insurance company)
Security Policy
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
• Essentially a statement of security requirements
• Every security policy statement should have a corresponding
enforcement mechanism
• Policies are at multiple levels
• High level policies flow down to multiple lower level policies
 High level; e. g., “company proprietary information shall be protected
from release to unauthorized personnel”
 Mid level; e. g., “there shall be no externally initiated ftp sessions”
 Low level; e. g., a firewall rule blocking incoming traffic on ports 20 (ftp
data), 21 (ftp control), and 69 (tftp)
 The firewall is the enforcement mechanism
• Policies also define management processes (e. g., incident
response actions) and personnel rules (e. g., don’t write down
passwords)
Security system engineering
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
• Part of overall systems engineering process
• Iterates requirements, design, review through multiple
levels of detail
• Includes design and development
• Lower level security policies developed
• Security should be an integral element from the start
Student talks
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
• Presentations will focus on management and processes, not
technical details (you know them already)
• Presenter will be given basic references and other reference
pointers, and is encouraged to search for more material
• Presenter to assign background reading the week before the talk
• Review presentation with me for guidance as you develop it
• Prepare for ~ 45 minutes of presentation material, but use one
hour+ with discussion
• Active participation of audience is encouraged
• Presenter to assign homework on topic
Course Schedule (tentative)
•
•
•
•
1/26
2/2
2/9
2/16
• 2/23
• 3/2
• 3/9
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
*Overview (Hery)
*Risk Analysis (Hery)
*Secure Systems Engineering (Hery)
ISO 17799 (taken)
SSE/CMM (secure syst. eng. maturity model)
Policy (2 hours???--Hery)
Legal and other requirements
Security Management and administration of
Deployed Systems (2 hours)
Incident Response
Business Continuity Planning (merge w/ above?)
Course Schedule (tentative) continued
• 3/16*Assessment/Assurance (Hery)
*Architecture of Classified Systems (Hery)
• 3/30Security Engineering for Software
TRANSEC/EMSEC/Tempest (EE background)
• 4/6 Physical Security/tamper resistance
Information System Security Officer
• 4/13Government Key Management Policy
Security Audit (tentatively taken)
• 4/20Certification and Accreditation
Ethical issues in system design/management??
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
Student Team Project
Qui ckTime™ and a TIFF (Uncompressed) decompressor are needed to see thi s picture.
• Teams of ~3 students
• Pick a system (discuss choice with me)
 Want simple functionality, security issues, whole system (e. g., client
and server side)
•
•
•
•
•
Submit a 1-2 page proposal to management (Dr. Hery)
Assess risks, threats, vulnerabilities
Develop a security policy
Do a high level system security design
Present a “preliminary design review” (PDR) to management
(include risk analysis, policies, system architecture)
• Iterate on risk assessment, policy, design
• Present a final “critical design review” (CDR) to management and
the class
• Write a final report to management on above