Security Metrics What Can We Measure
Security Metrics What Can We Measure
Security Metrics What Can We Measure
Zed Abbadi
The Public Company Accounting
Oversight Board
What is a “Metric”
A metric is a system of
related measures enabling
quantification of some
characteristic. A measure is
a dimension compared
against a standard.*
Security metric is a system
of related dimensions
(compared against a
standard) enabling
quantification of the degree
of freedom from possibility
of suffering damage or loss
from malicious attack.*
*QoP’ 06, Oct. 2006
Do We Really Need Metrics?
"If you cannot measure it, you cannot
improve it."
"In physical science the first essential step in the
direction of learning any subject is to find
principles of numerical reckoning and practicable
methods for measuring some quality connected
with it. I often say that when you can measure Lord Kelvin
what you are speaking about, and express it in
numbers, you know something about it; but when
you cannot measure it, when you cannot express
it in numbers, your knowledge is of a meagre and
unsatisfactory kind; it may be the beginning of
knowledge, but you have scarcely in your
thoughts advanced to the state of Science,
whatever the matter may be." [PLA, vol. 1,
"Electrical Units of Measurement", 1883-05-03]
“You cannot manage what you cannot measure"
Drivers For Metrics
Money matters
Asset vs. liability
Governance
You claim it is a
science?
Do as good as math,
physics and astrology
Decision aid
How are we doing with
security
Good Metrics. vs. Metrics
Quantitative
Objective
Based on a formal model
Has a time dimension
Universally acceptable
Has ground truth
Inexpensive
Obtainable
Repeatable
Data Collection
Vulnerabilities,
exploits and attacks
Organization vs.
industry vs. everyone
else
Disclosure Policies
Accuracy
Statistical Significance
Attempts at Measuring Security
TCSEC (Orange book)
ITSEC (Europe’s
Orange book)
CTCPEC (Canada’s
Orange book)
Common Criteria
(everyone’s Orange
book)
Framework rather than
a list of requirements
SSE-CMM
NIST FIPS-140 series
NIST SP 800-55
Security Metrics Types
Process Security Metrics
Network Security Metrics
Software Security Metrics
People Security Metrics
Other
Process Security Metrics
Measure processes and procedures
Imply high utility of security
policies and processes
Relationship between metrics and
level of security not clearly defined
Compliance/Governance driven
Generally support better security
Actual impact hard to define
Examples
No. of Policy Violations
% of systems with formal risk
assessments
% of system with tested security
controls
% of weak passwords (non-
compliant)
No. of identified risks and their
severity
% of systems with contingency plans
Network Security Metrics
Driven by products
(firewalls, IDS etc)
Readily available
Widely used
Gives sense of
control
Nice charts and
interfaces
Can be misleading
Examples
Successful/unsuccessful
logons
No. of incidents
No. of viruses blocked
No. of patches applied
No. of spam blocked
No. of virus infections
No. of port probes
Traffic analysis
Software Security Metrics
Software measures are
troublesome (LOC, FPs,
Complexity etc)
“Laws of Physics” are
missing
Metrics are context
sensitive and
environment-dependent
Architecture dependent
Aggregation may not
lead to strength
Examples
Size and complexity
Defects/LOC
Defects (severity,
type) over time
Cost per defect
Attack surface (# of
interfaces)
Layers of security
Design Flaws
People Security Metrics
Relevance
Unique
characteristics
Risk perception
skewed “optimism
Bias”
Limited memory and
attention span
Behavior modeling is
difficult
Awareness training?
Reliability vs. Security
Similar but different
We care more about
reliability
Different adversary
model
Reliability models
exist, but…
Security is a moving
target
Most Common Security Metric
Risk- We love this thing!
Source for profit
Where is the data?
Non monetary consequences
Adversary behavior models
Accuracy against ground truth
Mission/system/support models
Dynamic in nature
Future Of Security Metrics
Consumers demand better
security metrics
Government involvement is
increased
Science evolves to provide
better measures
Vendors volunteer (forced
to) develop universal accurate
metrics
Some vendors cheat, a
watchdog is created
Security problems continue,
no change in level of risk