The following is a partial list of the projects I've worked on during my time with the Experimental Systems Group here at CCS:
Tenwen was the first real systems project I was a part of, after I'd graduated from the basic software installs that the crew has traditionally started with. Tenwen was basically a complete redesign, from scratch, of CCS' computing environment. Remy Evard, the leader of Systems, authored a paper describing Tenwen, which he presented at the Eigth Large Installation Systems Administration Conference.
The knowledge I gained while working on Tenwen (starting just 5 short months after I first logged into a UNIX machine) has been the foundation of all the work I have done since as a member of Systems.
One of the very first projects Remy assigned me when I first became a member of systems was building a set of standardised kernels for our SPARC workstations running SunOS 4.1.3. As it happened, I barely knew what a kernel was at the time, but that's hardly unusal for a systems project.
With lots of time spent with SunOS manuals, and a lot of help from Ivan Judson, I built a kernel "sandbox" directory for kernel builds, and built standard kernels for our sun4c and sun4m machines, as well as some special kernels (including multicast patches). Probably the most time-consuming piece of this project was installing all of the OS patches from Sun. When I was through, we had a comprehensive system for building, documenting, and installing new kernels for our SPARCs, and scripts to do much of the above.
The goal of this current project is to update the administrative end of the ccs.neu.edu zone of the Domain Name Space. As our network has grown, so has the load on our single local nameserver, and it's time we create secondaries. We're also going to run locally-compiled BIND-4.9.X, rather than whatever hacked-up versions of BIND our vendors chose to ship with their operating systems.
CCS recently received 22 DEC ALPHA workstations as part of a research grant. These machines will primarily be used for visualisation and high-speed networking projects. At the moment, they're simply on our regular ethernet, but they'll soon be on their own CDDI ring(s). (That's FDDI over copper instead of fiber).
I really have done more than this during my tenure on systems, but we tend to suffer from a lack of documentation, which is why there aren't more nice project pages to link to. Of late, there's been a new effort to document, since it's so important, so hopefully more pages will be forthcoming soon.