Welcome!

I'm Daniel, a computer geek. Online I usually go by the nicknames of Dunje (often stylized as Dunj3), or Kingdread in the past.

I'm interested in most things computer and technology related. I've been a Linux user for quite some years, and I administrate some Linux servers for various purposes. I also like to program, mostly working on (personal) open-source projects. I work as a researcher in the field of (IT) privacy and security.

Aside from computers, I enjoy bicycling, photography, video games and some chess.

Software Projects

My main programming languages are Rust and Python, although I have also worked with C, C++, Java, Scala and JavaScript before. I am also looking for excuses to brush up on my Haskell skills, though I have not yet been very successful.

My open-source projects are on my GitHub, GitLab or Codeberg accounts. I also mirror (most of) them to git.kingdread.de.

My (biased, non-objective) highlights are:

Research

I am a researcher at Karlsruhe Institute of Technology, working at the Chair of Privacy and Security. My focus are Anonymous Communication Networks such as Tor or Mixnets. There I aim to analyze existing systems, try to find attacks or weaknesses, or develop improvements for those systems.

In addition to standard anonymous communication, I also explore more exotic forms, such as anonymous communication for groups or in ad-hoc networks.

Publications

If you’d like to read a publication that you cannot access, get in touch with me per e-mail.

Errata / Addenda

On re-reading our PolySphinx paper, it seems like we were a bit careless to explain who reconstructs the message plaintext: In the paper, we provide the final mix node with the seed for the key tree and the path. However, that means that a malicious final hop can compare the various intermediate ciphertexts with the ciphertexts that arrived at other (malicious) nodes to potentially re-trace the message. This is not what we intended.

Instead, our design relies on the fact that we trust the message recipients (as they are group members), and we can therefore reveal this information to the recipient (but not the final hop). Therefore, the seed and path information should be encrypted for the recipient, and then used locally by them to recover the plaintext. Alternatively, the onion can be extended by one layer to include the recipient as the last hop (but then we don’t need to include the recipient).

Miscellaneous

fietsboek.org: Website of fietsboek, including more documentation and screenshots.

Contact

You can contact me via the following channels: