Spark-based Application for Abnormal Log Detection

Ericsson is a world-leader in the rapidly-changing environment of communications technology and thus it is important to provide reliable and high quality networks. Automated test loops are executed frequently, trying to find problems  in Ericsson's products but, since test cases alone are not a...

Full description

Bibliographic Details
Main Author: Koutsoumpakis, Georgios
Format: Others
Language:English
Published: Uppsala universitet, Institutionen för informationsteknologi 2014
Online Access:http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-233335
id ndltd-UPSALLA1-oai-DiVA.org-uu-233335
record_format oai_dc
spelling ndltd-UPSALLA1-oai-DiVA.org-uu-2333352014-10-03T06:46:38ZSpark-based Application for Abnormal Log DetectionengKoutsoumpakis, GeorgiosUppsala universitet, Institutionen för informationsteknologi2014Ericsson is a world-leader in the rapidly-changing environment of communications technology and thus it is important to provide reliable and high quality networks. Automated test loops are executed frequently, trying to find problems  in Ericsson's products but, since test cases alone are not always adequate,  machine learning techniques  are sometimes  used to find abnormal system behaviour. The Awesome Automatic Log Analysis Application (AALAA) tries to find such behaviour by checking the log files produced during the testing, using machine learning techniques. Unfortunately,  its performance is not sufficient as it requires  a lot of time to process the logs and to train the model. This thesis manages to improve AALAAs performance by implementing a new version that uses Apache Spark, a general purpose  cluster  computing system, for both the processing of the data and for the training of the machine learning algorithm. Student thesisinfo:eu-repo/semantics/bachelorThesistexthttp://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-233335IT ; 14 057application/pdfinfo:eu-repo/semantics/openAccess
collection NDLTD
language English
format Others
sources NDLTD
description Ericsson is a world-leader in the rapidly-changing environment of communications technology and thus it is important to provide reliable and high quality networks. Automated test loops are executed frequently, trying to find problems  in Ericsson's products but, since test cases alone are not always adequate,  machine learning techniques  are sometimes  used to find abnormal system behaviour. The Awesome Automatic Log Analysis Application (AALAA) tries to find such behaviour by checking the log files produced during the testing, using machine learning techniques. Unfortunately,  its performance is not sufficient as it requires  a lot of time to process the logs and to train the model. This thesis manages to improve AALAAs performance by implementing a new version that uses Apache Spark, a general purpose  cluster  computing system, for both the processing of the data and for the training of the machine learning algorithm.
author Koutsoumpakis, Georgios
spellingShingle Koutsoumpakis, Georgios
Spark-based Application for Abnormal Log Detection
author_facet Koutsoumpakis, Georgios
author_sort Koutsoumpakis, Georgios
title Spark-based Application for Abnormal Log Detection
title_short Spark-based Application for Abnormal Log Detection
title_full Spark-based Application for Abnormal Log Detection
title_fullStr Spark-based Application for Abnormal Log Detection
title_full_unstemmed Spark-based Application for Abnormal Log Detection
title_sort spark-based application for abnormal log detection
publisher Uppsala universitet, Institutionen för informationsteknologi
publishDate 2014
url http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-233335
work_keys_str_mv AT koutsoumpakisgeorgios sparkbasedapplicationforabnormallogdetection
_version_ 1716716052694433792