Monitoring the energy consumed by a network infrastructure to detect and isolate faults in communication architecture

TL;DR

This paper presents a method for monitoring and analyzing energy consumption in network infrastructure using SDN to detect device state changes and anomalies, aiming to improve energy efficiency and fault detection.

Contribution

It introduces a novel approach combining real power data with power models to detect and isolate faults in network devices within SDN-managed infrastructures.

Findings

Effective detection of device state changes

Successful identification of energy anomalies

Potential integration with SDN controllers

Abstract

In recent years, a number of major improvements were introduced in the area of computer networks, energy-efficient network protocols and network management systems. Software Defined Networking (SDN) as a new paradigm for managing complex networks brings a significant opportunity to reduce the energy consumption among ICT. In this paper, we are tackling improvements in the process of monitoring the states of the networking devices and optimizing the existing solutions. We are monitoring the energy consumption of a network architecture and augment the retrieved raw power data to detect changes in the state of the devices. The goal is to benchmark the difference between the power data fetched from the real-measures and the data extracted from the power models, translated as the expected behavior of the devices. An application is designed to monitor and analyze the retrieved power data of a simulated ICT infrastructure composed of Cisco switches and routers, Dell Precision stations and Raritan PDUs. Moreover, smart algorithms are developed which outcome results with detection of changes in the state of the devices, as well as with detection and isolation of possible anomalies and their impact on the energy consumption. The application is envisioned to be an extension to the existing SDN controllers for monitoring and reporting the changes in the state of the ICT devices.