Big Data Architectures For Robust Systems

Falk, Eric

Reference : Big Data Architectures For Robust Systems


	Document type :	Dissertations and theses : Doctoral thesis
	Discipline(s) :	Engineering, computing & technology : Computer science
	Focus Areas :	Computational Sciences
	To cite this reference:	http://hdl.handle.net/10993/34596

Title :	Big Data Architectures For Robust Systems
Language :	English
Author, co-author :	Falk, Eric [University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > >]
Publication date :	15-Nov-2017
Institution :	University of Luxembourg, Luxembourg, Luxembourg
Name of the degree :	Docteur en Informatique
Number of pages :	150
President of the jury :	Voos, Holger
Secretary :	Ormazabal, Gaston
Member of the jury :	State, Radu
	Engel, Thomas
	Gurbani, Vijay
Keywords :	[en] Big Data Architectures ; Streaming Data Analytics
Abstract :	[en] Monitoring robustness of critical systems/infrastructures is the major use case for anomaly detection. A robust system designates a structure not only safe against intentional attacks, but also capable of stemming internal failures. These systems face two primary risks: cyber attacks fall into the first category, whereas failing hardware components are part of the second category. In both cases, fast decision making is crucial. Hence, streaming data processing is the decisive asset to consider. With this background, in this thesis, we investigate two scenarios from the fields of mobile network sanity monitoring and cyber-physical security. Our contribution is threefold: We display how the real-time requirements of the two use cases push existing frameworks to their utter limits; We show which anomaly detection methods can be used to facilitate instant assessment rendering; We blueprint the extensions we contributed to big data frameworks, which are powering major silicon valley companies, to make them capable of supporting our use cases. The data-sets issued by our monitoring systems yield different properties than data from internet companies such as Google, Facebook or LinkedIn. In this work we establish our use cases, illustrate the mathematical models employed for the decision taking, and examine how big data architectures have to be altered to support our scenarios.
Target :	Researchers ; Professionals ; Students ; General public
Permalink :	http://hdl.handle.net/10993/34596

File(s) associated to this reference

Fulltext file(s):

	File	Commentary	Version	Size	Access
Limited access	dissertation.pdf		Author postprint	5.72 MB	Request a copy

All documents in ORBi^lu are protected by a user license.

University of Luxembourg Library | Feedback | Legal notices