Buffer Overflows & Security Policies

• These are study notes based on a sample exam.

Buffer Overflows

• A buffer overflow happens when a buffer receives more data than it can store, overwriting adjacent memory locations.
• Key elements include the absence of bounds checking on memory buffers and the use of low-level memory access, such as in C/C++.
• Types include stack overflows, heap overflows, and format string attacks.
• Exploitation involves input exceeding buffer size, overwriting the return address with a pointer to malicious code, leading to the execution of attacker's code upon return.
• Prevention strategies are input validation and bounds checking, utilizing bounds-checked functions like strncpy, compiler protections like stack canaries, DEP, ASLR, and employing memory-safe languages like Java or Python.

Information Security Policies & Digital Signatures

• An information security policy constitutes a formal document outlining rules for information protection, covering access control, acceptable use, password rules, and incident response.
• An example includes a password policy mandating strong passwords and regular rotation.
• Distinctions between policy and procedure lie in the "what" (policy) versus the "how" (procedure).
• Digital signatures ensure message integrity and authentication through the use of private/public keys.
• The process involves hashing a message, encrypting the hash with the sender's private key to form a signature and the receiver then hashes the message, decrypts the signature using the sender's public key, and confirms integrity if the hashes match.
• Weaknesses include private key theft and reliance on a trusted Public Key Infrastructure (PKI).

Programming Language Safety & Type Soundness

• Memory safety prevents access to memory outside valid bounds.
• Type safety ensures variables are used consistently with their declared type.
• Type soundness guarantees type rules are maintained during execution.
• Unsafe features include pointer arithmetic (in C/C++) addressed by bounds-checked structures, and manual memory management that can be improved with garbage collection.
• An example of a C++ type violation is casting an object to access private fields, like reinterpreting an object's memory as a long to directly set a value.
• Java offers protections: no pointer arithmetic, array bounds checks, and automatic garbage collection.

Abstractions & Least Privilege

• Layered abstraction involves separation between hardware, OS, language runtime, and applications.
• Each layer enforces boundaries and policies.
• Example failures include a stack overflow crossing into the OS level or a VM escape compromising the host OS.
• Least privilege dictates that users and software should have only the minimum permissions required.
• Examples provided include non-root web server user accounts, Role-Based Access Control (RBAC) within applications, and code sandboxing techniques.
• Each abstraction layer enforces limited access to lower layers.

Threats & Threat Modeling

• A threat represents a potential event or actor that can exploit vulnerabilities.
• Threat modeling identifies and prioritizes security threats in a system.
• STRIDE is a model for classifying threats: Spoofing, Tampering, Repudiation, Information Disclosure, Denial of Service, Elevation of Privilege.
• The threat modeling process involves identifying assets, diagramming architecture, decomposing the application, identifying threats using models like STRIDE, documenting threats and mitigations, and rating threats using DREAD.
• Example countermeasures: HTTPS for eavesdropping, input validation for injection attacks, and strong authentication for spoofing.
• DREAD model components: Damage, Reproducibility, Exploitability, Affected users, Discoverability.