I will be assisted Prof. Koopman with the double checking of errors in cyclic redundancy checks and checksums in network messages. The objective is to understand how efficient error detecting codes are at identifying defective messages and seeing if a probability can be placed on certain methods of error checking.
To accomplish these goals, messages with randomized asserted bits will be generated and have correctly calculated CRCs or checksums appended to them. These messages will then be corrupted with error bits and tested to see if the errors are detected. The messages will differ in length and error detection code.
Due to the magnitude of possible errors/messages that can exist, it would be difficult to test every possible case. Instead, an extensive amount of messages will be tested and analyzed.
Results will differ with the form of error detection code used, size of message and amount of error bits inserted. From this data, a probability value may be able to be calculated.