User login

Jayden Hewer's blog




This week I redesigned my sack test to selectively acknowledge multiple blocks. which has resulted in the start and fail states being more clearly defined. It can now with a degree of accuracy determine if sack behaviour is being demonstrated by the server or not.

Along with added multiple sack blocks I also worked on resolving edge cases which showed up during testing.




This week I worked on my sack amp test, which requires a basic implementation of TCP to handshake and send a HTTP request so it has data to then send a selective acknowledgement. Got it to the stage where it can handle expected behavior.

I also spent time giving my sack test the ability to craft packets with data and with a larger range of options. This was done to be able to send HTTP requests and the sack option.




This week I tested my second AMP test and ironed out the remaining issues and bugs. Because all the options where set in my TCP ack packet, if it was dropped due to one of the options I wouldn’t know why so I added a fall back which would test options individually if a few preconditions where meet. I also separated my tests into their own files to make it easier to see what was going on and make extensions easier as well as a generally refactoring of the code.




This week I worked towards getting an amp test complete that will test a list of TCP options to a location, seeing which options are accepted or if the packet is dropped all together. With the idea being that if this was done from many points in the network you could infer if a middle box is tampering with the options.