What I ended up doing this summer
I was a little bit embarassed to admit this, so I didn’t post earlier: I didn’t end up getting a job with Shotgun. For reasons not under my control, they were unable to hire me this summer. I’m still on good terms, and hopefully someday I’ll get a chance to work with them again; but not this summer.
Now that I’ve found some other work, I’m more comfortable with posting about it. I applied to and was accepted for a research position at UOIT, under the supervision of Dr. Heydari. I’ll be working with Dr. Heydari to model and simulate WAN networks, and more specifically the way that they fail. The concept will be that I will be simulating link and node failures, and watching what the network does in response. At first I will be using standard IP link failure handling, which is basically the stuff we learned in CCNA; routing protocols notice the change, and figure out the new best path, and then eventually converge. This takes between seconds and minutes. The goal is to make a huge improvement on this time within autonomous networks, possibly via a protocol called MPLS.
I’ll be working in an industry-standard network modeller and simulator known as OPNET. Part of my duties as a research student will be to study the capabilities of OPNET and find a way to fulfil the goals of the project. My first goals are to determine how to make a node or link disconnect, and furthermore, see if it is possible to disable all of the links within a certain radius of a coordinate (which will simulate severe a local outage, such as a natural disaster). The second goal that I have in the short term is to work on automation. I need to be able to get OPNET to load a topology, break the network in some way, log what happens, and dump that into a file for later review - all done thousands of times, in slightly different ways.
After that, I’m not sure. There’s a possibility might have to write some OPNET code (which is really just C with an OPNET library), in order to help with the modelling of MPLS and the improvement of speed of convergence. I’ll post more when I know.
