Sayed, Shereef2015-07-182015-07-182015-07-17vt_gsexam:5107http://hdl.handle.net/10919/54566As the complexity of software increases, so does the complexity of software testing. This challenge is especially true for modern military communications as radio functionality becomes more digital than analog. The Software Communications Architecture was introduced to manage the increased complexity of software radios. But the challenge of testing software radios still remains. A common methodology of software testing is the unit test. However, unit testing of software assumes that the software under test can be decomposed into its fundamental units of work. The intention of such decomposition is to simplify the problem of identifying the set of test cases needed to demonstrate correct behavior. In practice, large software efforts can rarely be decomposed in simple and obvious ways. In this paper, we introduce the fuzzing methodology of software testing as it applies to software radios. Fuzzing is a methodology that acts only on the inputs of a system and iteratively generates new test cases in order to identify points of failure in the system under test. The REDHAWK implementation of the Software Communications Architecture is employed as the system under test by a fuzzing framework called Peach. Fuzz testing of REDHAWK identified a software bug within the Core Framework, along with a systemic flaw that leaves the system in an invalid state and open to malicious use. It is recommended that a form of Fault Detection be integrated into REDHAWK for collocated processes at a minimum, and distributed processes at best, in order to provide a more fault tolerant system.ETDIn CopyrightSecurityFuzzingSCAOSSIEREDHAWKThesis