Move over Zoolander, this is not about weather that upsets your hair-do.
Threat modelling is a depiction of critical security concerns. Sometimes represented diagrammatically as a Data Flow Diagram illustrating potential external attack points.Why you do this exercise is that a threat model can help to assess the potential harm that an attacks will bring and help us anticipate thus minimise the risks.
It is in essence a set of approaches that you use to allow you to identify potential threats and conduct this as new systems are being designed. This helps you understand your actual vulnerabilities.
Understand my vulnerabilities
By doing threat modelling for both IT infrastructure and IT applications we are able to really understand the risks that cyber threats bring. In other words we develop an understanding of the expected attack surface.
This is critical that you look at the bigger picture as it is very easy to fall into the trap of focusing on operational patching and alerts. Every organisation has a long list of identified risks with specific vulnerabilities that are at various stages of being remediated.
Any organisations that are able to link their threat modelling and enterprise risk analysis will have a much better understanding of the cyber risks. This clear understanding of the attack surface will then enable me to prioritise my cyber security tasks.
This needs to be based on real actual risks not perceived risk, or even worse hype that is currently manifesting.
Approaches to Threat Modelling
There is no perfect way to conduct threat modelling. But here three well regarded approaches:
Privacy approach – this is an asset-centric threat modelling, where you look at where personally sensitive data is being stored, transmitted etc. Just focus on the asset and follow it through the business.
Microsoft approach – this is Software-centric threat modelling. You look at the architecture, commencing with the design of the system and walk through evaluating threats against each component.
Black Hat – start from the outside in. By looking at what the attacker is trying to do. This is often called an Attacker-centric threat model.
Getting your act together
At an Enterprise level, the worst situation is that there is a department that conducts risk management and a separate division that does threat modelling.
Having no interaction really will provide for your business the worst of both worlds.
From my experience, the typical risk management committee reviews risks every month and reports these to the risk board meeting. The fact that this is intermittent and when the world is working in real time continuously is an issue.
The intervention that you as a leader need to do is to create active link between risk management and threat modelling. An example of the benefit is that then after a penetration test is completed and we learn about new vulnerabilities then this is linked to our risk management. This just means we get to root causes, rather than just having meaningless activity.
Cyber Insurance to drive Threat Modelling
I think what is going to happen is that we see more and more enterprises adopting Cyber Insurance. But as insurers will face difficulty in pricing risk then they will start to asset that organisations demonstrate that they are following good practices around threat modeling and risk management.
As high profile security issues continue to get attention, we will need to ensure that we avoid the ‘beauty’ contest and really look hard at the essence of the issue at hand.