Random walks and chemical networks

We consider continuous-time random walks with bounded jumps but unbounded rates (they depend polynomially on coordinates in the orthant). In applications, the case when the rates are bounded corresponds in applications to queueing theory, more precisely, to Markovian communication networks. The goal of this paper is to discuss the situation for polynomial rates; we show that the boundaries often play no role, but new effects and complicated behaviour may arise due to time scales and nonlinearity.