An SNMP Based Traffic Characterisation Paradigm for Green-Aware Networks

An SNMP Based Traffic Characterisation Paradigm for Green-Aware Networks

Kiran Voderhobli
DOI: 10.4018/978-1-4666-8210-8.ch014
OnDemand:
(Individual Chapters)
Available
$37.50
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

This chapter describes a novel approach to study network patterns in a data centre with the aim of reducing power consumed. Cloud infrastructures rely on numerous networked devices in data-centers to provide virtualization and sharing of resources. Network traffic is one of the key contributors to power consumption. Numerous techniques to develop power-aware data-centers have been proposed in the recent years. Virtualization management is based on many critical decisions including work-load, utilization, location of physical resources etc. This chapter takes a unique network management angle to greening a data center. It describes how Simple Network Management Protocol (SNMP) has a great potential to characterize traffic which can then feed into decisions for management of virtualized entities.
Chapter Preview
Top

Introduction

Affordable Business Computing

The business model of subscription based computing has brought affordable IT services into the realms of everyone. Cloud computing has proved to be a very cost-effective means of sourcing IT, even for smaller firms and some third-world countries, where availability of powerful hardware and software resources are an expensive affair (Avram, 2014; Marston et al, 2011). Since cloud computing is considered as a business model, the cloud based resources are available in various profiles like private, public, community and hybrid clouds to address the requirements of various classes of users (Gupta, Seetharaman & Rudolph Raj, 2013). The cloud model allows for provisioning of Infrastructure as a Service (IaaS) which involves allocation of virtualised resources on demand. This could be software or specialist hardware like networking equipment, data store etc. Software as a Service (SaaS) gives users access to software via the Internet. This could be as simple as a word processor or as complex as any full-scale ERP software. Platform as a Service (PaaS) offers services related to backend platforms, middle-ware, SDKs and frameworks. More recently, there has also been focus on specific business related functionalities that could be delivered via the cloud. For example, Business Intelligence as a Service (BIaaS) delivers services like dataset analysis, financial analysis, stock analytics, business trends analysis etc. (Chang, 2014). Such services can help businesses off-load their “number crunching” tasks to the capabilities of cloud computing. The above features of the cloud are strong motivations for businesses to embrace cloud based business computing (Ramachandran & Chang, 2014). The business advantages combined with the inherently green property of the cloud, is another strong incentive to adopt cloud computing.

Cloud Network Operations and Energy

In the modern business computing landscape, cloud based IT infrastructure seems to be the norm. The low cost of computing and ease of access to powerful hardware and software resources is the rationale for adopting cloud based computing solutions. Cloud computing has revolutionised the way applications and services are delivered to businesses and end-users. The innovation of the cloud has helped mask the technicality of powerful computing resources from end-users. Subscribing to cloud based services alleviates the problem of having to build an IT infrastructure from scratch. In the last few years, the number of companies using cloud based infrastructures has constantly been on a rise due to Green policies widely adopted. Governments and corporate bodies have started to push for IT policies that have lower carbon footprints. Since cloud computing allows for sharing of resources, it reduces the need to deploy more hardware, thereby reducing power consumption. But traditional sustainability policies have only considered high level operations rather than looking at how data processing and data flow could contribute to efficiency in green computing. Network operations arising from virtualisation contribute to power consumption. It is therefore important to consider network operations in a cloud infrastructure to optimise operations that will help reduce energy consumption. One key aspect that is often overlooked is the behaviour of packet transmission and network activity. Management operations of virtualisation should be based on continuous evaluation of network activity as this could be vital to saving power. There must be a body of knowledge that will help save power when there really is a scope to do so. In order to derive this body of knowledge, traffic patterns of virtualised instances (like a VM) must be studied.

Complete Chapter List

Search this Book:
Reset