I'm Chris, a research scientist at the Pittsburgh Supercomputing Center, part of Carnegie Mellon University. I've been working there for 25 years and have focused my career on making networks faster, easier to use, and more secure.

One of my most important projects is HPN-SSH. Based on OpenSSH this is focused on providing the fastest possible network throughput to securely move large datasets. Work on HPN-SSH started almost 15 years ago and initial successes saw a performance improvement of 2 orders of magnitude in some circumstances. This improvement was based on making SSH aware of the current size of the TCP receive buffer in autotuning kernels. Other advances have been the development of a threaded AES-CTR cipher, allowing users to switch to the NONE cipher after authentication (for non-sensitive data transfers), and other other performance improvements.

Recently I've begun work on automatically resuming failed transfers in SCP, improving the performance of the ChaCha20 cipher, upgrading the AES-CTR cipher to use on die AES instructions, reducing memcpy overhead by normalizing I/O buffers, and the like. Our long term goal is to make HPNSSH as fast and as secure as possible in order to support the growing data and network needs of researchers, businesses, and average users.

Sponsors would support this work by providing resources to hire student developers, build out of community engagement efforts, deploy necessary test hardware, and fund further development and maintenance of HPN-SSH.