• Koldehofe, Boris
• Saleh, Saad

Abstract

The current internet architecture comprises of a series of data processing entities, called packet processors, requiring high energy consumption and lacking cognitive functionalities for the Internet traffic. With the discovery of a novel physical component Memristor, the network packet processors can provide cognitive functions with less energy consumption than the prior state-of-the-art techniques. The combination of storage and computation inside the memristors can efficiently deploy neuromorphic computing models on network packet processors along with a decrease in the energy consumption and latency. In this research, we present the cognitive models & network functions using the memristor-based novel computing architectures for use in network packet processors.

Topics

• laser emission spectroscopy