Challenges of Massive Concurrency

Work at the interface of Computing and Biology has to confront problems in massive concurrency, where by "massive" I mean as in the Law of Mass Action, where there are order of Avogadro number of concurrent interacting processes, which are rarely if ever present in single copy. These systems have been traditionally, and still often appropriately, modeled as deterministic dynamical systems, even though the stochastic/discrete nature of molecular interactions has been much emphasized recently, bringing the topic closer to Computer Science. In simple chemical reactions, agents (molecules) are stateless and have simple interactions, but the picture is completely different in biochemistry where even a single agent (protein molecule) can easily have an internal state space of 2^12 and dozens of interaction channels. These systems form, for example, sophisticated information processing networks with a highly combinatorial and typically unbounded state spaces, and again should be of intrinsic interest to Computer Science. The technically distinguishing feature in these systems is the inherently continuous nature of time (either deterministic time or stochastic time), which separates it from much, but not all, of the process algebra and verification literature. I will illustrate these issues with a couple of case studies in DNA computing and Systems Biology where we attack problems, mostly separately, with discrete and continuous methods.