On Memristors for Enabling Energy Efficient and Enhanced Cognitive Network Functions

Abstract

The high performance requirements of nowadays computer networks are limiting their ability to support important requirements of the future. Two important properties essential in assuring cost-efficient computer networks and supporting new challenging network scenarios are operating energy efficient and supporting cognitive computational models. These requirements are hard to fulfill without challenging the current architecture behind network packet processing elements such as routers and switches. Notably, these are currently dominated by the use of traditional transistor-based components. In this article, we contribute with an in-depth analysis of alternative architectural design decisions to improve the energy footprint and computational capabilities of future network packet processors by shifting from transistor-based components to a novel component named Memristor. A memristor is a computational component characterized by non-volatile operations on a physical state, mostly represented in form of (electrical) resistance. Its state can be read or altered by input signals, e.g. electrical pulses, where the future state always depends on the past state. Unlike in traditional von Neumann architectures, the principles behind memristors impose that memory operations and computations are inherently colocated. In combination with the non-volatility, this allows to build memristors at nanoscale size and significantly reduce the energy consumption. At the same time, memristors appear to be highly suitable to model cognitive functionality due to the state dependence transitions in the memristor. In cognitive architectures, our survey contributes to the study of memristorbased Ternary Content Addressable Memory (TCAM) used for storage of cognitive rules inside packet processors. Moreover, we analyze the memristor-based novel cognitive computational architectures built upon self-learning capabilities by harnessing from non-volatility and state-based response of memristors (including reconfigurable architectures, reservoir computation architectures, neural network architectures and neuromorphic computing architectures).

Nonvolatile Memristive Materials and Physical Modeling for In‐Memory and In‐Sensor Computing

“…Thus, a substantial improvement in the computing efficiency is achieved when the boundary between the memory and the processing module becomes negligible, which mimics an energy efficient human brain where the processing and memory elements are interlinked. [24][25][26] Memristive materials (we term them MMs) are promising candidates for achieving in-memory computing. [27][28][29][30][31] Upon the application of an external electrical stimulus, the MM system discloses programmable conductance states.…”

“…Thus, a substantial improvement in the computing efficiency is achieved when the boundary between the memory and the processing module becomes negligible, which mimics an energy efficient human brain where the processing and memory elements are interlinked. [ 24–26 ]…”

LiNbO3-based memristors for neuromorphic computing applications: a review

“…The ferroelectric nature, electro-optic properties, nonvolatility, low power consumption, high-speed switching, and the capability for multilevel resistance states collectively position LiNbO 3 as a multifaceted material for developing efficient and brain-inspired computing architectures. Accordingly, the exploration of LiNbO 3 in neuromorphic computing stems from its potential to mimic synaptic behavior and facilitate energy-efficient computing processes (Saleh and Koldehofe, 2022;Xu et al, 2023).…”

Synaptic Properties of Geopolymer Memristors: Synaptic Plasticity, Spike-Rate-Dependent Plasticity, and Spike-Timing-Dependent Plasticity

“…Memristors are solid-state devices characterized by nonlinear electrical properties and tunable resistance that allow them to mimic the synaptic plasticity of the brain and make them suitable for the development of neuromorphic devices. , Memristors are compatible with traditional CMOS technology, and they do not suffer from the von Neumann bottleneck limitation. , Non-von Neumann computing architectures based on memristors are composed of artificial synapses and neurons that perform both data processing and memory functions via in-memory computing. , The collocation of memory and processing units reduces the computational complexity of the computing devices. Additionally, using memristive synapses in non-von Neumann computing devices result in energy-efficient and enhanced cognitive network functions . The idea of memristors was postulated by Chua in 1971, and concretized into a solid-state TiO 2 device in 2008 by the HP lab. , The memory abilities of memristors are usually observed as a pinched hysteresis in the IV (i.e., current–voltage) plot, – which is the result of the history-dependent conductance change of the switching material.…”

“…Additionally, using memristive synapses in non-von Neumann computing devices result in energy-efficient and enhanced cognitive network functions. 13 The idea of memristors was postulated by Chua in 1971, 14 and concretized into a solid-state TiO 2 device in 2008 by the HP lab. 15,16 The memory abilities of memristors are usually observed as a pinched hysteresis in the IV (i.e., current−voltage) plot, 17−19 which is the result of the historydependent conductance change of the switching material.…”