Choosing a topic for a dissertation (postgraduate study)

Question

IT education

Entered graduate school - the question arose of choosing a topic for a dissertation. Advise any that are of interest to you - will be considered. If worthy developments appear, I will write on Habr.

What comes to the mind of the patient:
- homomorphic encryption + cloud computing;
anything related to clouds at all;
— data transfer protocols (peer-to-peer, encryption, media transfer);
— programming languages ​​/ their extension (AOP library for .NET);
- ...

In general - any useful direction.

Answer 1

Network Function Virtualisation

Answer 2

For P2P, it is now relevant to study peer search algorithms in the perspective of organizing online viewing of torrents, because. there is an opinion that massive sequence_downloads will greatly worsen the quality of distributions.
Why do I think this is relevant? It would seem that with the growth of channels , this topic has lost its meaning, because an “average” movie is downloaded in 20 minutes. But this is an erroneous impression: P2P is already migrating to I2P, which throws us back “many years ago” in terms of speeds.
Therefore, it would be very useful to develop a methodology for assessing the quality of distributions, and then find out in which specific cases online viewing noticeably worsens this very “quality”.
As a practical result, there may be an implementation of an online browsing algorithm that “does not interfere” with the distribution as a whole, for example, in our open source P2P client (which we have been making for many years, about 1 million users). With a serious attitude to the matter, we guarantee comprehensive assistance and support from our full-time programmers. If interested, details in a personal. :)
PS My downvoted question on this topic .

Answer 3

So what was the diploma about? very many simply continue the theme of the graduation project.

Answer 4

The current topic (and interesting in my opinion) is the detection of anomalies in traffic, intrusion detection, protection against network attacks. You can try neural networks and genetic algorithms in applications to different areas of technology - voice recognition, images, logistics. If you are not afraid of mathematics, you can try to describe some information processes (the same search for anomalies in traffic) from the standpoint of catastrophe theory (I have seen several publications on this topic). I myself was going to write an article on this topic on Habré, since my scientific work, although from a different area, intersects in terms of the mathematical apparatus.

Answer 5

In the volumes of D. Knuth "The Art of Programming" there are many tasks of varying complexity, among which there are still unsolved problems (of high complexity). There are many of them and, accordingly, there are many different themes.