An international team of scientists led by Germany's Ruhr-Universität Bochum (RUB) is developing new techniques to efficiently identify coding errors in embedded systems via a system called Fuzzware.
The researchers use fuzzing algorithms to feed random inputs to software to try to crash the application.
They expedite the fuzzing process by narrowing down possible inputs, which RUB's Thorsten Holz said involves employing only those inputs "that the firmware expects and can handle."
Holz explained the process resembles the Infinite Monkey Theorem in that the fuzzer eventually, by chance, will use meaningful inputs after enough attempts. The team tested 77 firmwares with Fuzzware, checking up to three times more code than conventional fuzzing methods in the same amount of time.
From Ruhr-Universität Bochum (Germany)
View Full Article
Abstracts Copyright © 2022 SmithBucklin, Washington, DC, USA
No entries found