In the future, we can expect swarms of robots to autonomously perform useful work for us, but to do so effectively the individual robots must be able to communicate with one another and with outside parties. While the simple assumption for many years has been that communications are a solved problem, it has been shown time and again that communications are not a solved problem, and much research remains to be done. In this talk, Cem Karan surveys a number of approaches that have been researched to control robots to maintain communications. The talk is thematically organized through the lens of the CAP (Consistency, Availability, and Partition-tolerance) theorem, a theorem that comes from database theory that can be used to explain why communications problems persist despite decades of research. A few selected papers are presented in greater detail, but the majority are provided as references for the interested listener to research on their own. This talk is expected to last approximately 50 minutes, with 10 minutes for questions afterwards.