The stability issue in a k-star routing network with permanent edges and a token-based scheduling policy is discussed in this paper. The model is derived from the notion of adversarial queueing theory. The authors first prove the general conditions in which a finite token-full network becomes stable, and then explain, in detail, the architecture of the k-star network. Several conditions leading to either stability or divergence are suggested and verified.
The network architecture investigated in this paper is characterized by token passing scheduling and star topology. Both were of research interest between the late 1980s and early 1990s. Aside from legacy systems and specialized configurations, they are not the prevailing architectural options these days. Therefore, the paper’s contributions are somewhat weakened. That said, this paper should be very helpful to those who are interested in network performance modeling, and queueing theory in general. The authors did a fine job of describing the problem and presenting their proofs. I certainly enjoyed reading the paper. The modeling approach and techniques may inspire similar works in other networking environments.