Runtime measure.docx - Run-Time Countermeasures Executable...

Run-Time Countermeasures
Executable Space Protection – A lot of exploits build on injecting and executing malicious code
• By separating the memory space of a process into executable and modifiable parts, code
injection can be prevented. Limitation: cannot fully protect programs that create and execute
code at runtime – NXU bit
Counter-countermeasure Circumventing Executable Space Protection – The attacker can re-use
existing code from the memory space of the process for malicious purposes. The attacker has
control over the stack → attacker can set up parameters for the C library function
Address Space Layout Randomization (ASLR) • r andomly arrange the positions of the
executable, the stack, and the heap in the process’s address space • may prevent return-to-libc
attacks • most operating systems (e.g., Windows, Linux) implement some randomization •
Counter-countermeasures • information leakage (e.g., print vulnerability) • random guessing.
Information Leakage with printf Vulnerability . can be used to figure out the address of the
executable can be used to figure out the address of the stack. Limitations of ASLR • stack and
heap cannot be located at any address. Heap spraying • fills up the memory with a certain
sequence of bytes.
Secure Programming - Trade-off problem: secure programming can increase development time
(and cost), and decrease the performance of the software product.
Input Validation Approaches – Blacklist - list “known bad” inputs • do not allow inputs that
are on the blacklist. listing all bad inputs is difficult and error-prone • typically lower impact on
usability. Whitelist - list “known good” inputs • allow only inputs that are on the whitelist.