RNS Applications in Computer Networks

Abstract

Residue number system (RNS) is a non-weighted number system, which represents any integer number with some residues, based on the selected moduli set. Therefore, computations can be done on these residues in parallel to decrease the delay, and also due to the reduction of bit-length, low power processors can be designed. Moreover, another characteristic of this number system is its ability to improve the reliability by using redundant moduli. These are the most famous advantages of RNS, which numerous works have been done on them, and they led to valuable achievements [1]. However, recently modern applications of RNS have been introduced which profit from RNS in new aspects with distinct goals and also new challenges [2]. RNS can be used in network applications with the aim of solving some issues such as determining output port in simpler way [3] or increasing a wireless sensor network (WSN) lifetime [4]. Moreover, it is possible to design a switch which can process multicast or unicast traffic by using RNS concepts [3,5]. Redundant residue number system is even used in cloud storage with the aim of providing a novel method to enhance reliability [6]. Another application of RNS in WSN is its usage in increasing the reliability of data transmission by improving packet reception rate and end-to-end delay [7]. Furthermore, RNS can be applied to WSN in order to perform data aggregation [8] which results in a good trade-off between energy consumption and reliability.This chapter briefly reviews some of the important and effective applications of RNS in computer networks to show how RNS can improve performance in different ways than conventional methods. In other words, our aim is creating a

A popularity-based caching strategy for improved efficiency in SVRF-based multicast control-planes

“…Error-free membership query schemes, SVRF [32] and its derivative-fractional-N SVRF [33] were proposed to overcome the disadvantages of BF. SVRF constructs and queries group memberships based on utilizing RNS properties including the continuous product (CP) and Chinese remainder theorem (CRT) [34], [35] to improve unicast and multicast packet forwarding in a PFE. It traverses the PFE and encodes the entire forwarding table to a 'scalar-pair', which represents a single forwarding entry as a 'vector'.…”