Hypothesis
Quantum resistant cryptography algorithms can be as secure and fast as current RSA and Diffie-Hellman based algorithms. This will result in an economical transition once new standards are set by the appropriate governing bodies.
Research Approach
Research on this subject will be primarily conducted using quantitative data collected from a variety of algorithms. This data will come from existing studies and information released by the algorithm authors. Experimental data will also be collected from current implmenetations and compared to classical algorithms on common hardware.
Measurements
The best method of measurement is a direct comparison. For experimental data collected, the time to encrypt and decrypt common datasets will be compared. Other tradeoffs from data obtained from studies will be compared such as key and handshake sizes.
Analysis
Analyze the suitability, strengths and weaknesses of each algorithm. As this is an extremely new field of study a broad overview of findings and suggestions will be presented.
Footnote
This has proven to be a difficult blog post to put together and I think the research proposal will be tricky. I’m not confident that even the best mathematicians in the world could come up with a concrete conclusion to the hypothesis because of the sheer number of unknowns in this new field of study (however it’s generally accepted to be true). This would result in more of an informational study for others to base their work off than one that comes up with any new conclusion.
While this is a subject I’ve been very interested in and have been following for a while, I think a more well understood subject that could result in a clear conclusion from experimental data like many others I’ve read would have been far easier to write about.
Yep, that does seem too broad. You are welcome to go in a 180-degree turn. If you’ve found any studies that interested you that you can easily propose replicating, press the easy button and do that instead.
When I suggest to students that they change topics, they often fear they’ll need to redo all previous assignments. That is NOT what I’m suggesting. But, research is a small step forward and, with luck, a smaller step back.
If you started with this idea and have run it through only to realize it won’t work well, don’t become Sisyphus, pivot to a new direction.
I will also note that when you compare Hank and Aki’s proposed study to the original one they based their proposal on, I think you’ll see that the original had a number of additional dependent variables. As they worked on the research, I strongly advised them to cut those down. The fewer variables you are considering, the fewer variables you need to explain.
Yesterday I started putting together a research proposal and found some benchmarking studies very similar to what I was thinking. There’s not very many of them but the ones I found were very well put together.
I’m thinking of rewording the hypothesis a bit, but leaving the general theme in place. The study will mainly focus on collecting experimental data (along with a bit of non experimental data such as key sizes) and making direct comparisons on them. Since the general consensus is a fast algorithm with small key size that has a favorable relationship between security and key size is ideal.
This would give me something concrete to write a conclusion about. I think that sounds a lot more workable than what I wrote in this blog post.