Risk analysis is a complex science. Hopefully, the following notes may provide a framework upon which to study this in more depth:
WHAT IS RISK ANALYSIS?
A TEXT BOOK DEFINITION
A procedure to identify threats & vulnerabilities, analyze them to ascertain the
exposures, and highlight how the impact can be eliminated or reduced.
ANOTHER DEFINITION
A process to
determine what security is appropriate for a system or environment.
THE BOTTOM LINE
The security
you implement should be commensurate with the risks applicable. Risk Analysis
should enable you to achieve this goal.
It should
also help you establish where to invest your security budget for the best
return.
TYPES OF RISK ANALYSIS
QUANTITATIVE RISK ANALYSIS
-
Two Elements Used: Probability and Likely Loss
-
An 'ALE' is produced (via Probability x Likely Loss)
-
Several Drawbacks
Eg: no accurate probability database
probability is usually unique to case
expected loss hard to establish
'expected' not easy to accept!
- In
fairly limited use
QUALITATIVE RISK ANALYSIS
- Widely used
-
Estimated potential loss/impact used
-
No probability database required
-
Risk 'level' often produced
QUALITATIVE RISK ANALYSIS
MEASURING RISK
BUILDING A SECURITY MODEL
ELEMENTS IN THE EQUATION
THREAT: Nasty
things that can happen
ATTACK: Made
by a threat when it occurs
VULNERABILITY: Weakness... makes a system more prone to
attack or an attack more likely to succeed
CONTROL: A control
is a 'countermeasure' for a vulnerability.
IMPACT: A
successful attack has an impact
BUSINESS
IMPACT: This is what we must
reduce or prevent!
ALL
THESE PUT TOGETHER GIVE US A FRAMEWORK TO MANAGE SECURITY WITH
THREAT/VULNERABILITY/CONTROL
EXAMPLES
THREAT
Fire
Software Error
VULNERABILITY
Presence of
Complexity
flammable
materials
CONTROLS
Sprinklers
Design and
Extinguishers
development
Etc
standards
Change control
QUALITATIVE RISK ANALYSIS
PRACTICAL APPLICATION
USING THIS MODEL WE CAN BEGIN TO MEASURE AND QUANTIFY.
FOR INSTANCE: VULNERABILITY v CONTROL
A
significant vulnerability with no or little control in place to address it is
bound to increase the risk of a successful attack.
Conversely,
low vulnerability and substantial control will reduce the risk of successful
attack.
HOWEVER
If
the IMPACT is always going to be LOW, then substantial control may NOT be
necessary or desirable, and may signify overspend.
This
is one of the reasons why ALL the elements in the model must be integrated into
your risk analysis approach (some techniques do not provide full linkage with
impact).
THIS
IS NOT TRIVIAL TO ACHIEVE... BUT IF YOU CAN ACHIEVE IT, THE BENEFITS ARE
SUBSTANTIAL

|