srieger1/digsinet

AFLNet experiments

srieger1 opened this issue · 2 comments

Check and describe:

  • assumption a: as soon as one crash occurs, multiple others quickly follow (due to branch handling in aflnet process), hypothesis: it takes (way) longer to wait for the first crash as it will take to wait for subsequent ones (x crashes in a row can be assumed to be a vulnerability, x = ?), characteristic increase / steps in diagram for multiple runs? similar lines?
  • assumption b: it takes around x secs (based on paths / sec) until a crash is found
  • assumption c: because of the genetic mutational approach of AFLNet, inputs that repeatedly prove successful in increasing the code coverage and still provoking unexpected behaviour get created. These inputs can then be used for more extensive testing and examination, thus greatly aiding in the security enhancement process.

@Lachstec I think we can close this issue for now, as our main focus is now on message queues and event-streaming

@srieger1 Yep! I'm fine with that. Closing this now.