Cyclic-Reservation Multiple-Access (CRMA) is an access scheme for high-speed local and metropolitan area networks based on a folded-bus or dual-bus configuration. CRMA provides high throughput and fairness independent of the network speed or distance. In CRMA, the headend generates reserve commands periodically. Each station may reserve a number of empty slots in each reserve command if necessary. Corresponding to every reserve command, the headend generates a cycle of length equal to the total number of slots reserved. Every cycle is used to serve the reservations made on its corresponding reserve command. Generally, a longer cycle length means a longer access delay and a lower throughput. Therefore, it is desirable to have a scheme to make the cycle length as short as possible. In this paper, we will study the problem of reducing the total number of empty slots generated within every cycle as much as possible. However, it has been shown that the problem is NP-complete under the constraint that all the slots used by a station in a cycle are required to be consecutive [1][N.F. Huang, H.I. Liu, G.K. Ma, A cycle compression algorithm for CRMA high-speed networks, IEEE ICC'93, 1993, pp. 1363–1367]. In this paper, we discover that if this slot-contiguity constraint is released, then the problem can be optimally solved in polynomial time. Furthermore, we propose an optimal slot scheduling algorithm with time complexity O ∑ i≔1 M α i 2 where α i is the number of empty slots reserved by Station i in the cycle and M is the number of stations. To evaluate the effectiveness of our algorithm, a large number of computer simulations were performed. We compare our algorithm with the original CRMA in terms of the following three important performance measurements: average cycle length, average throughput, and average medium access control (MAC) delay. Experimental results demonstrate that our algorithm has much shorter cycle lengths, much higher throughput, and much shorter MAC delay. Therefore, our algorithm is very attractive.
Read full abstract