Reliable multistage interconnection network design

Abstract

High-performance supercomputers generally comprise millions of CPUs in which interconnection networks play an important role to achieve high performance. New design paradigms of dynamic on-chip interconnection network involve a) topology b) synthesis, modeling and evaluation c) quality of service, fault tolerance and reliability d) routing procedures. To construct a dynamic highly fault tolerant interconnection networks requires more disjoint paths from each source-destination node pair at each stage and dynamic rerouting capability to use the various available paths effectively. Fast routing and rerouting strategy is needed to provide reliable performance on switch/link failures. This paper proposes two new architecture designs of fault tolerant interconnection networks named as reliable interconnection networks (RIN-1 and RIN-2). The proposed layouts are multipath multistage interconnection networks providing four disjoint paths for all the source-destination node pairs with dynamic rerouting capability. The designs can withstand switch failures in all the stages (including input and output stages) and provide more reliability. Reliability analysis of various MIN architectures is evaluated. On comparing the results with some existing MINs it is evident that the proposed designs provides higher reliability values and fault tolerance.

Reliability Analysis of Multistage Interconnection Networks

“…The present work evaluated reliability calculations on various MINs . Although the approaches are applied to a realistic network such as computer communication networks, telecommunication networks, and transport networks, complex networks are used to interconnect multiple processing units in distributed systems.…”

“…Redundant paths are all possible paths for any source–destination nodes pair, while disjoint paths are redundant paths that do not have any element in common other than source and destination nodes. Providing more disjoint paths enhances MIN fault tolerance, as the common elements leading to multiple path failures are absent.…”

“…This method generates the path sets of desired order of cardinality and obtaining the path sets in increasing order of their cardinality and finally evaluating reliability from ordered minimal path sets using MVI algorithm. MVI algorithm successfully calculated reliability values for all the tag values of GIN, ES‐GIN, 2DGINs, 3DGINs, and 4DGINs …”

“…Failure of nodes in any of this communication system affects its users adversely and shifts our focus on evaluating the effect of node failures on reliability evaluation. Reliability of an MIN is defined as the probability that an MIN will be able to provide the required interconnection for a specified period of time under a given set of operating conditions.…”

Design and Reliability Evaluation of Gamma-Minus Interconnection Network

“…Performance metrics of Gamma MIN and its variants reported in the literature may be analyzed from its fault tolerance capability, disjoint path-set, reliability, cost effectiveness, network complexity, etc. [34][35][36][37][38][39][40][41] Fault tolerance may be defined as the capability of the network to root the packet through an alternative/redundant path even in the presence of fault/switch-busy in primary path between any source-destination node pair. [11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29] So the idea behind increasing the fault tolerance capability of network is to provide more redundant paths in the network.…”