Know your neighbor: self‐organizing data streaming processing in Advanced Metering Infrastructure

Examensarbete för masterexamen

Please use this identifier to cite or link to this item: https://hdl.handle.net/20.500.12380/202649
Download file(s):
File Description SizeFormat 
202649.pdfFulltext1.02 MBAdobe PDFView/Open
Type: Examensarbete för masterexamen
Master Thesis
Title: Know your neighbor: self‐organizing data streaming processing in Advanced Metering Infrastructure
Authors: Kotsantis, Theodoros
Abstract: Devices in fields like telecommunication and infrastructure networks as well as telephone and stock markets are sources of continuous data streams. Data streams are usually data created by sensors. These sensors are great in number, for example the sensors that automatically measure electricity consumption in a country can be hundreds of thousands. Also these sensors can vary the rate they create new data. In addition several of the applications that create data streams require close to real time processing of data. For example, having access in real time to the electricity consumption of all the users in one district, we could detect if the part of the network that serves that district is getting overloaded and take on actions that would prevent an outage. So in order to process data streams, we should address three challenges: process a) high volume of data, b) with a fluctuating volume in a c) close to real time fashion. Stream Process Engines (SPEs) are the systems that emerged from the Database (DB) community and are designed to process data streams in an online fashion. Stream Engines process the data streams as soon as data arrive by applying continues queries. Each data element is processed at most once, without the need to persist the information first. Stream engines can process great volume of data with close to real time delay. Stream engines can run in a distributed fashion but we need to use Load Balancing (LB) protocols, so the stream engines would cope with fluctuations in the rate that data arrive. One system that could benefit from Stream Processing is the Advanced Metering Infrastructure (AMI), the system responsible for the reporting of end user electricity consumption to utility management office. For this thesis we assume on the AMI there is installed a distributed SPE application. We assume that the AMI that we examine in this thesis uses a hierarchical network to transfer information (i.e. smart meter readings). Distributed SPE systems that exist in bibliography don’t provide a LB protocol for a distributed SPE application with hierarchical structure. For this reason we want to develop a LB protocol that can apply to the distributed SPE application of our AMI system. In this thesis, we developed a load balancing protocol that can avoid overloading and can spread the workload among the entities of distributed SPE application. We studied the characteristic of such a system, developed the load balancing protocol, implemented and evaluate its performance in a simulated environment.
Keywords: Data- och informationsvetenskap;Computer and Information Science
Issue Date: 2014
Publisher: Chalmers tekniska högskola / Institutionen för data- och informationsteknik (Chalmers)
Chalmers University of Technology / Department of Computer Science and Engineering (Chalmers)
URI: https://hdl.handle.net/20.500.12380/202649
Collection:Examensarbeten för masterexamen // Master Theses



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.