Root Cause Analysis of Mobile Site Outage Using Bayesian Network: the Case of ethio telecom
No Thumbnail Available
Date
2018-11-17
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
AAU
Abstract
In most cellular networks Trouble Shooting (TS) is a manual process, accomplished
by Radio Access Network (RAN) experts. Their task is to resolve problems
in the network that have been identified by other employees or by automated
checking routines. During the TS procedure, several applications and databases
have to be queried to analyze performance indicators, cell configuration and
alarms of the cells. For example when an outage in mobile network investigated,
a Trouble Ticketing (TT) reflects the problem status, which is the fault description.
A TT system is deployed as a large database, which can be queried by the
user, using criteria like time constraints or identifiers, such as site ID (site identity).
After a query, all the cases related to the specified site within the given time
period are shown. The entries are normally in “free text”, almost like a “virtual
log-book” that everyone uses to annotate the actions taken and observations made
for the site. Hence, in difficult cases lots of people from various departments have
looked into the site and potentially applied some changes. For example a case
might involve changing parameters or swapping hardware. Thus, several notes
are normally written to such a case.
The availability of mobile services without interruption has many social and economic
benefits. However, mobile network outages occur due to many different
reasons. One of the reasons for Base Transceiver Station (BTS) site outage is failure
of hardware elements of the most varied kinds: switching units, cables, cooling
elements, energy elements, etc. The possible relations or interconnections between
elements are not explicitly well-defined so study the root causes of network outage
is important.
In telecom network environment, there are different network problems of which
their associated causes can be address through Artifitial Intelegent (AI), mostly
detection, and forecasting. On the other hand incident management systems (also called Trouble Ticket (TT)) have hardly used AI techniques to optimize the processes
involved. Thus this thesis work investigate the root causes of network outage
using Bayesian network models, then analyzing the ethio telecom (ET) network
outage TT data.
The outcome for the analysis result are model based Root Causes Analysis of
BTS mobile site network outage is performed, the network technicians can be informed
of the real scope of failures and the probable existence of root problems,
technicians can be advantages through managing the situations and good for decision
making task by providing the root causes of alarms and their probabilities
of occurrence.
Description
Keywords
Network outage, Bayesian Networks, Root cause analysis, Cellular mobile network, Trouble Ticketing, Probability of failure, Model