KCSA · Domain 6: Compliance and Security Frameworks · 10% of the exam
Task 6.2: Threat Modeling Frameworks
Structured ways to find threats in a system, applying them to a cluster's components and data flows.
Study it
Compliance, threat modeling and supply chain frameworks
Lesson coming
Sample questions
Try each one before opening the answer. Every option is explained, with the CNCF documentation page that proves it.
Question 1 · choose 1
A team has one workshop day to threat model a new payment service. Its draft diagram shows the service's Pods, Services, ingress and database. The code lives in a hosted Git service, images are built by a shared CI system and kept in a registry, and four teams own parts of the service. The team wants the day's output to be a threat list it can trust. Following the CNCF Cloud Native Security Whitepaper, how should the team set up the model before listing threats?
AKeep the runtime diagram and record the Git service, CI and registry as trusted external parties
BExtend the scoped model to the code location, storage and build path, mark boundaries, model crossings
CUse ATT&CK's threat matrix for containers in place of a system model, as it lists the known techniques
DLet each of the four teams model its own part of the service separately, then merge the lists
Show the answer and why
AKeep the runtime diagram and record the Git service, CI and registry as trusted external parties
Incorrect
The whitepaper says cloud native diagrams must go beyond the core runtime components to include where the source code lives, the storage in use and the parts of the development cycle; marking them trusted leaves them unexamined.
BExtend the scoped model to the code location, storage and build path, mark boundaries, model crossings
Correct
All threat modeling techniques start with a scoped representation of the architecture, its processes, data stores and security boundaries, then model how elements interact, with special attention to interactions that cross boundaries; for cloud native this scope includes the source code location, the storage and the development cycle.
CUse ATT&CK's threat matrix for containers in place of a system model, as it lists the known techniques
Incorrect
The matrix describes adversary behavior during an attack and helps to evaluate threats, but every technique the whitepaper describes starts from a scoped model of the team's own system.
DLet each of the four teams model its own part of the service separately, then merge the lists
Incorrect
Separate models miss the interactions between the parts, and the whitepaper asks for special attention to interactions that cross security boundaries.
The whitepaper's threat modeling starts with a scoped representation of the system: important processes, data stores and security boundaries, then the interactions between elements, especially those that cross boundaries. For cloud native systems the diagrams should also show where the source code lives, the storage mechanism and the development cycle, so a runtime-only diagram, a generic attack matrix or separately built partial models leave gaps in the threat list.
A cluster's threat model has one actor, the malicious outsider, because every engineer is vetted and RBAC authorizes each action they take. Then an engineer with deploy rights applies a manifest from a convincing phishing message, and it opens a backdoor. The security lead wants the model's actor list fixed so that controls are designed for this kind of event as well. Following the CNCF Cloud Native Security Whitepaper, what should change?
AAdd authorized insiders, both malicious and uninformed, as primary actors next to outsiders
BAdd the uninformed outsider as a primary actor, since the phishing message came from outside
CAdd the malicious insider only, since RBAC already limits what a deceived engineer can do
DKeep the outsider-only list and treat the incident as a training issue outside the model
Show the answer and why
AAdd authorized insiders, both malicious and uninformed, as primary actors next to outsiders
Correct
The whitepaper's primary actors are malicious insiders, uninformed insiders (authorized actors, assuming anyone can be duped) and malicious outsiders; the deceived engineer is an uninformed insider.
BAdd the uninformed outsider as a primary actor, since the phishing message came from outside
Incorrect
The engineer who acted holds deploy rights, so the actor is an insider; the whitepaper treats uninformed outsiders as optional, with controls likely a subset of those for the primary actors.
CAdd the malicious insider only, since RBAC already limits what a deceived engineer can do
Incorrect
RBAC authorized the engineer's harmful action; the whitepaper lists the uninformed insider as a primary actor of its own because anyone can be duped.
DKeep the outsider-only list and treat the incident as a training issue outside the model
Incorrect
Vetting does not take staff out of the model: the whitepaper's primary actors include authorized insiders, malicious and uninformed.
The whitepaper names three primary threat actors: malicious insiders, uninformed insiders and malicious outsiders; uninformed outsiders may be added for completeness, with controls that are likely a subset. An engineer with authorization who is deceived into a harmful action is an uninformed insider, and RBAC does not help, since the action was authorized.
For identifying threats to cloud native systems, which mature threat models does the CNCF Cloud Native Security Whitepaper name as examples?
ASTRIDE or OCTAVE
BCVSS or EPSS
CKerberos or LDAPS
DRBAC or ABAC
Show the answer and why
ASTRIDE or OCTAVE
Correct
The whitepaper recommends leveraging a mature, well-used model of threats such as STRIDE or OCTAVE.
BCVSS or EPSS
Incorrect
These score vulnerabilities; they are not threat models.
CKerberos or LDAPS
Incorrect
These are directory and authentication protocols, not threat models.
DRBAC or ABAC
Incorrect
These are authorization models, not threat models.
In its threat identification step, the whitepaper suggests a mature threat model such as STRIDE or OCTAVE and lists example threats for each STRIDE category in a cloud native setting.
In ATT&CK's threat matrix for containers, as described by the CNCF Cloud Native Security Whitepaper, what do the rows and columns represent?
ARows are tactics and columns are techniques
BRows are CVEs and columns are affected images
CRows are techniques and columns are tactics
DRows are Kubernetes components and columns are versions
Show the answer and why
ARows are tactics and columns are techniques
Incorrect
The whitepaper describes it the other way round.
BRows are CVEs and columns are affected images
Incorrect
The matrix describes adversary behavior, not vulnerability inventories.
CRows are techniques and columns are tactics
Correct
The whitepaper says the threat matrix has rows and columns, where the rows are the techniques and the columns are the tactics.
DRows are Kubernetes components and columns are versions
Incorrect
The matrix is organized by adversary tactics and techniques.
ATT&CK's threat matrix for containers lays out adversary tactics, the attacker's goals such as initial access, as columns, and the techniques used to reach them as rows.