- Mental models (how to decide) are necessary for effective Cyber Security (not just what to buy or how to implement).
- Cyber Security without strategies is sub-optimal (I will explain with a Sun Tze Art of War strategem 謀攻).
- Attack Surface management is a priority & something within your control because it will make you look at things from both perspectives: Offense & Defense.
A mental model is simply a representation of how something works. We cannot keep all the details of the world in our brains, so we use models to simplify the complex into understandable and organizable chunks. - https://fs.blog/mental-models/
What is Cyber Defense? How is it different from Cyber Security? Cyber Defense is about the strategy (of what matters & why) employed to disrupt attacks, while Cyber Security focuses on how (tactically) to protect your network from threats. Strategies & Tactics are two sides of the same coin. Let’s dissect Cyber Defense with a First Principle mentality.
Strategy without tactics is the slowest route to victory. Tactics without strategy is the noise before defeat - Sun Tze
There is no shortage of models & guides related to “HOW” you should implement products, controls & compliance with checklists. For instance, NIST’s “Identify, Protect, Detect, Respond & Recover” functions, controls for “Authentication, Authorization & Accounting”, Information Security requirements of “Confidentiality, Integrity & Availability”.
From Attack by Stratagem (謀攻) chapter:
Translates (by Lionel Giles) to:
Hence the saying: If you know the enemy and know yourself, you need not fear the result of a hundred battles. If you know yourself but not the enemy, for every victory gained, you will also suffer a defeat. If you know neither the enemy nor yourself, you will succumb in every battle.
Most high-level frameworks & controls help with “knowing yourself” (indirectly) in terms of networks & assets. Many companies (& even government agencies) appeared in headlines for the wrong reasons related to breaches are pretty well-equipped with compliance checklists & controls.
Mistranslation of this stratagem assumes “victory when we know our enemy & ourselves”, which it is never the case for Cyber Defense since most organisations do not retaliate. We can at best work to remain “undefeated”, which is what “不殆” means. Unenviable job because no one remembers defenders’ “victories”, but everyone remembers the breaches.
Sadly, many are in the category of “know neither the enemy nor themselves”. They have poor network visibility & situation awareness that attackers can literally take their own sweet time to explore the victim’s network.
With these in mind, let’s examine the problem at its roots.
Fire fighters & law enforcers don’t start off learning about fire engines, water pumps, patrol cars & fire-arms. Instead, every FireFighter needs to know the Combustion Triangle & for Law Enforcers, they learn Motivation, Means & Opportunity when evaluating suspects. So why should Cyber Defenders start off with assuming “prescriptions of HOW to ...” (largely in form of buying products & ticking checkboxes) would assure safety & security?
For (most) fire to burn, we need three ingredients to start a fire. Some may argue that exceptions like phosphorus self-ignites at room temperature without the application of heat, but that’s just initial reaction, to sustain any combustion, these 3 ingredients are still necessary.
Similarly, offenders (Threat Actors) have the Motive (for most cases, unless we discount insanity), Means & Opportunity (Routine Activity Theory - Cohen & Felson, 1979) to act on victims (you & your networks).
Giving criminals an opportunity to snatch your valuables turns out to be a lot like oxygen for fire to burn. So what are the necessary & sufficient conditions of any Cyber Attacks?
Three tenets for secure cyber-physical system design and assessment are a threat-driven quantitative mathematical framework for secure cyber-physical system design and assessment. This empirical approach has been used by the US Air Force Research Laboratory (AFRL) for secure system research and development. The Tenets were first documented in 2005 as a teachable methodology.
The Three Tenets threat model posits that three ingredients are necessary and sufficient for cyber-physical vulnerabilities to exist: i) system susceptibility, ii) threat accessibility, and iii) threat capability. Just like fires & crimes, all ingredients need to be present. Fortunately, like the earlier animations, we can avert problems by removing or disrupting one or more of the conditions.
- A system is susceptible when it has value (to attackers) & vulnerabilities (useful for attackers).
- Attackers need access to know about system susceptibilities, thus accessibility to the system’s Attack Surface, which is the sum of access points to deliver attacks, either physically or logically.
- Even with access & knowledge, Threat Actors must possess the capabilities in terms of tools, techniques & resources in order to succeed.
Strategy #1: cutting off “oxygen” denies opportunities for attackers. So which of the three tenets is the “oxygen” & why (next section)?
Fuel (prices are high now)
Cyber attacks are prevalent simply because all three ingredients are always present. Threat Accessibility is the only condition that we can fully control because:
- Most systems, regardless of custom-built or bought off-the-shelve or cloud, are flawed.
- Adversaries have the advantage of freely available offensive tools & knowledge.
- It is costly to defend in terms of effort & shortage of staff.
- Legacy controls & mindsets (or ignorance) persist.
The above are unfortunate (non-exhaustive) facts to live with, BUT we can choose HOW we expose our assets to networks.
To put the first mental model (i.e. the attack conditions) to practice (& experience why Threat Accessibility is the “oxygen”), all my interns (get bored just reading the 3-tenets paper) start off with this first lesson on Attack Surface management:
- Sign up free cloud Virtual Machines running Linux with SSH (remote access) service & ZeroTier Virtual Private Network (this is NOT a sponsored post & ZeroTier is more of a Software-Defined Perimeter for micro-segmentation than traditional VPN which is a gate-to-whole-internal-network).
- Learn which log & events to look at to observe persistent brute-force SSH authentication attempts (Threat Capability) immediately after starting the VM instance (before joining VPN).
- Shut off SSH service quickly to join VPN (a private network that consists of their laptops & the server VM).
- Realise despite poor password (system susceptibility), a simple configuration limits the service to only accept connections from within the VPN, thus denies all brute-force attempts (just like the earlier fire blanket demo, server log no longer shows any new events after the configuration).
This exercise does not imply VPN services are the silver bullet & systems will become impenetrable, or we can be lazy with basic hygiene & condone weak authentication. But the odds of remaining “undefeated” is much BETTER when you manage Attack Surface as a priority because:
- It is often quickest, least effort & impact, especially for administrative services (privileged access thus high-value targets);
- It buys you time to prioritise your patching, starting from assets that are directly exposed, to the next hop & so on. As oppose to the naïve approach of “crown jewels” (first because your spreadsheet says so) that are far away from the initial pivot & often high impact when patching screws up;
- Attackers think in terms of connected-graphs & not a list with arbitrary sorting (yes, those silly “risk” formulas in your spreadsheets) based on opinions or preferences!
So let’s conclude this first part of three by relating the necessary & sufficient conditions with Attack Stratagem:
Know your enemies
Keeping up with system patches (needs to be done by simply not enough because it is not just exploitable bugs that are useful for attackers, abusing system features is a prevalent tactic to evade controls & detection)
Subscribe to Threat “Intel” stories (there are better ways to spend your budget)
Executing Attack Surface & Vulnerability Management plans that make sense for your use-cases, networks & budget
- Reconnaissance (gather information)
- Infiltrate Networks (aka Initial Access)
- Spread (within networks aka Lateral Movement)
- Execute Objectives (that impact Safety of Cyber-Physical Systems; Confidentiality, Integrity & Availability of Information-Systems)
Attackers tend to RINSE (mnemonic) & repeat around the necessary & sufficient attack conditions.