I had a lengthy meeting with Ian today to debrief my plan for the coming weeks and iron out some doubts I had about both the project itself and the dissertation.
According to Ian, I shouldn’t worry about expanding the literature review from my proposal yet, but keep it as-is, and add to it if/when I need to refer to new papers to help me making progress on practical work.
When designing the packet system, I wasn’t sure wether I should align packet boundaries on the start of new frames, or just treat the packet stream as a continuous stream of binary data and have some packets start in the middle of a frame. According to Ian, the bandwidth saved by having a continuous stream is not worth the increase in complexity. So each packet will start with a new frame, and will then spread over the next N frames:
:::plain ______________________ ____________________________ | packet 1 | | packet 2 | |____________________| |__________________________| v____________ ______v______ v____________ _____________v | frame 1 | | frame 2 | | frame 3 | | frame 4 | |___________| |___________| |___________| |____________|
Ian confirmed that using FEC as a means to get a better overall bandwidth was not a good idea, as I’d suspected last semester when I started working on packet design.
Ian advised me to keep the hardware design part of the project for the end, if there is enough time left, to make sure I could get the important parts of the project (Computer Science) done properly first.