Abstract

AbstractThe Shared buffer memory switch (SBMS) architecture was originally proposed as an effective approach to implement ATM switch fabrics. However, in this paper we find that if an error occurs in the address chain memory of one linked list which stores the address of the next cell in the shared buffer memory, the erroneous situation will spread over all linked lists in the SBMS in a short time. In order to prevent the fault spread phenomenon, we propose two doubly linked list based architectures to combat address chain failure; these are referred to as the Flush and In‐Seq schemes. The first scheme flushes the remaining cells in the faulty queue but collect their addresses for later usage. The second scheme outputs the remaining cells in their correct sequence. From our simulation, if the error injection rate is low, the performance of the In‐Seq scheme experiences slight degradation compared with the errorfree situation.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call