Our current understanding of brain function is still too limited to take advantage of the computational power of even the simplest biological nervous systems. To fill this gap, the Si elegans project (www.si-elegans.eu) aims at developing a computational framework that will replicate the nervous system and rich behavior of the nematode Caenorhabditis elegans, a tiny worm with just 302 neurons. One key element of this emulation testbed is an electro-optical, micromirrorbased connectome. Unlike any other current ICT communication protocol, we expect it to accurately mimic the parallel information transfer between neurons. This strategy promises to give new insights into the nature of two hypothesized key mechanisms - the parallel and precisely timed information flow - that make brains excel von-Neumann-type machines. In this contribution, we briefly introduce the overall Si elegans concept to then describe the requirements for designing a light-based connectome within the given boundary conditions imposed by the hardware infrastructure it will be integrated into.