D4 - Statement of Research Questions.lyx

#LyX 1.4.3 created this file. For more info see http://www.lyx.org/
\lyxformat 245
\begin_document
\begin_header
\textclass article
\language english
\inputencoding auto
\fontscheme pslatex
\graphics default
\paperfontsize default
\spacing single
\papersize default
\use_geometry true
\use_amsmath 1
\cite_engine basic
\use_bibtopic false
\paperorientation portrait
\leftmargin 3cm
\topmargin 3.5cm
\rightmargin 3cm
\bottommargin 3.5cm
\secnumdepth 3
\tocdepth 3
\paragraph_separation indent
\defskip medskip
\quotes_language english
\papercolumns 1
\papersides 1
\paperpagestyle default
\tracking_changes false
\output_changes false
\end_header

\begin_body

\begin_layout Description
Name: Zhe Wei Chong
\end_layout

\begin_layout Description
Student\InsetSpace \space{}
Number: 40899013
\end_layout

\begin_layout Description
Supervisor: Marius Portmann
\end_layout

\begin_layout Description
Project\InsetSpace \space{}
Title: Analyzing and Extending a Wireless Mesh Network Monitoring
 System
\begin_inset VSpace medskip
\end_inset


\end_layout

\begin_layout Section*
Research Question
\end_layout

\begin_layout Standard
The main question for this research topic is how to best monitor a wireless
 mesh network for system and traffic-related statistics, and then present
 the data in a coherent way so that the overall health of the mesh may be
 determined.
\end_layout

\begin_layout Paragraph
(42 words)
\end_layout

\begin_layout Section*
Introduction
\end_layout

\begin_layout List
\labelwidthstring MM

\series bold
1.
 Understand the wider context and importance of the project
\end_layout

\begin_deeper
\begin_layout List
\labelwidthstring MM

\series bold
a.
 State the general topic and make a claim about why it is important.
\end_layout

\begin_deeper
\begin_layout Standard
Wireless mesh networking is a dynamically self-organized and self-healing
 network architecture involving multiple wireless nodes and routers interconnect
ed in an ad-hoc manner 
\begin_inset LatexCommand \cite{Akyildiz2005}

\end_inset

.
 There is a growing need for resilient and efficient networking technologies
 that can be deployed in environments where networking infrastructure does
 not exist.
\end_layout

\end_deeper
\begin_layout List
\labelwidthstring MM

\series bold
b.
 Describe what is generally known about this topic.
\end_layout

\begin_deeper
\begin_layout Standard
To determine the operational status of a wireless mesh network, active monitorin
g must be done to capture statistics from each mesh node and aggregate them
 into meaningful information.
 Judging from flunctuations in the statstics the 
\begin_inset Quotes eld
\end_inset

health
\begin_inset Quotes erd
\end_inset

 of a wireless mesh network and whether the routing protocols and its supporting
 infrastructure and other design factors has been effective enough to sustain
 the mesh can be determined.
\end_layout

\begin_layout Standard
There is usually no centralized network backbone present, such as 802.11
 routers, that mesh nodes may use to forward networking traffic.
 Therefore specialized routing protocols are used to forward traffic through
 neighbouring nodes towards the desired destination, for example 
\emph on
Ad-hoc on-demand distance vector
\emph default
 routing (AODV) and 
\emph on
Dynamic source routing
\emph default
 (DSR) 
\begin_inset LatexCommand \cite{1223819}

\end_inset

.
\end_layout

\end_deeper
\end_deeper
\begin_layout List
\labelwidthstring MM

\series bold
2.
 Summarize previous research
\end_layout

\begin_deeper
\begin_layout List
\labelwidthstring MM

\series bold
a.
 State the core ideas in the literature and structure them in a logical
 sequence.
\end_layout

\begin_deeper
\begin_layout Standard
There are several proposals for gathering management data in mobile networks.
\end_layout

\begin_layout Standard
The most common approach using existing standards would be using the 
\emph on
Simple Network Management Protocol
\emph default
 (SNMP) 
\begin_inset LatexCommand \cite{Henderson2006}

\end_inset

.
 A 
\emph on
manager
\emph default
 is responsible for polling a number of 
\emph on
SNMP agents
\emph default
 within nodes present on the mesh for useful variables regarding the node's
 condition.
 To achieve almost real-time monitoring, the poll interval must be frequent
 enough.
 However, this results in additional traffic within the mesh that may overload
 the network and impact its performace.
\end_layout

\begin_layout Standard
To avoid this, Chen at al.
 proposed the 
\emph on

\begin_inset Quotes els
\end_inset

Ad-hoc Network Management Protocol
\emph default

\begin_inset Quotes ers
\end_inset

 (ANMP), an architecture that clusters nodes hierarchically to reduce the
 amount of traffic directly between manager and agents 
\begin_inset LatexCommand \cite{Chen1999}

\end_inset

.
 Shen at al.
 identifies that in a wireless mesh network consisting of highly mobile
 nodes there needs to be a framework which can better adapt to dynamically
 changing network conditions 
\begin_inset LatexCommand \cite{Shen2003}

\end_inset

.
\end_layout

\begin_layout Standard
Ramachandran et al.
 proposes a different approach that takes advantage of less mobile yet more
 capable nodes for pervasive monitoring 
\begin_inset LatexCommand \cite{Ramachandran2004}

\end_inset

.
 The 
\emph on
Distributed Architecture for MONitoring mobile networks
\emph default
 (DAMON) introduces two unique concepts, categorising management data in
 relation to temporal significance, and 
\emph on
data sinks
\emph default
 for aggregating information.
\end_layout

\begin_layout Standard
The findings from these experiments is extensively documented by Kiese and
 Mauve 
\begin_inset LatexCommand \cite{1223819}

\end_inset

.
\end_layout

\end_deeper
\begin_layout List
\labelwidthstring MM

\series bold
b.
 Draw conclusions from the literature review.
\end_layout

\begin_deeper
\begin_layout Standard
Compared to wireless mesh routing protocols and design however, research
 on wireless monitoring and visualization is still quite scarce.
\end_layout

\end_deeper
\end_deeper
\begin_layout List
\labelwidthstring MM

\series bold
3.
 Prepare for the current research
\end_layout

\begin_deeper
\begin_layout List
\labelwidthstring MM

\series bold
a.
 List the gaps.
\end_layout

\begin_deeper
\begin_layout Standard
There is no sufficient research on multiple channel wireless mesh networks.
\end_layout

\begin_layout Standard
It has not been documented how much does management architecture affect
 the health of the mesh, and how well does a distributed approach in monitoring
 fare compared to conventional methods, e.g.\InsetSpace \space{}
1 SNMP manager with many agents.
\end_layout

\begin_layout Standard
It is also not well documented as how link measurements from SNMP agents
 can be collected and interpreted to determine the health of the mesh as
 there are no standardized benchmark settings for conducting experiments.
\end_layout

\begin_layout Standard
There are not many publicly available tools for aggregating collected monitoring
 data in wireless mesh networks and visualizing them in a coherent interface.
\end_layout

\end_deeper
\begin_layout List
\labelwidthstring MM

\series bold
b.
 List possible methods for addressing the gaps.
\end_layout

\begin_deeper
\begin_layout Standard
The current implementation of AODV by National ICT Australia (NICTA) can
 be analyzed to determine its efficiency, with improvements made so it can
 faster adapt to changes in link availability.
\end_layout

\begin_layout Standard
The effect of monitoring on a wireless mesh network may be studied by adjusting
 the variables on a live mesh testbed and then documenting the results.
\end_layout

\begin_layout Standard
Improvements may be made to the mesh software to support extended data.
\end_layout

\begin_layout Standard
A modular software stack that can be deployed on wireless mesh testbeds
 may be designed.
\end_layout

\end_deeper
\begin_layout List
\labelwidthstring MM

\series bold
c.
 Select a gap and a methodogy for addressing it.
\end_layout

\begin_deeper
\begin_layout Standard
The gap selected is the lack of tools that can be used for monitoring and
 visualizing wireless mesh networks.
 As such, a software package will be developed by taking multiple publicly
 available tools offering a subset of the features required and combining
 them into a stack that can be deployed at a wireless mesh testbed.
\end_layout

\end_deeper
\end_deeper
\begin_layout List
\labelwidthstring MM

\series bold
4.
 Research plan
\end_layout

\begin_deeper
\begin_layout List
\labelwidthstring MM

\series bold
a.
 State the overall goals and specific aims of the research
\end_layout

\begin_deeper
\begin_layout Standard
The aim of the research is to develop an intuitive interface to NICTA's
 wireless mesh monitoring system.
 This involves building modules for collecting monitoring data, aggregating
 into meaningful graphs, and presenting them coherently.
 With it the network topology, overall health and traffic flow of the wireless
 mesh can be observed, which allows the underlying design of the mesh to
 be evaluated more easily.
\end_layout

\begin_layout Standard
The data collected will include generic measurements such as data throughput,
 packet loss, response times and routing, as well as specific 802.11 information,
 such as link quality, signal strength and noise.
\end_layout

\begin_layout Standard
Studies will also be carried out by applying different approaches to the
 wireless mesh monitoring system while observing the overall status of the
 mesh.
 It should then be determined whether it is possible to significantly improve
 data dissemination within the mesh using a decentralized approach compared
 to the conventional methods.
\end_layout

\end_deeper
\begin_layout List
\labelwidthstring MM

\series bold
b.
 Outline the methods to be followed
\end_layout

\begin_deeper
\begin_layout Itemize
The interface to the wireless mesh network monitoring system will be achieved
 using the following methods:
\end_layout

\begin_deeper
\begin_layout Itemize
Replacing the original modules by NICTA so that it can be easily maintained
 in the future.
 This includes rewriting the SNMP scripts and replacing the network topology
 generation module with one that is more capable of meeting our needs.
\end_layout

\begin_layout Itemize
Extending the monitoring system to support reading custom measurement attributes
 via SNMP and implementing SNMP traps so that link failures can be quickly
 identified and visualized as such.
\end_layout

\begin_layout Itemize
Designing an interactive interface that can incorporate multiple graphs.
\end_layout

\end_deeper
\begin_layout Itemize
When a wireless mesh testbed has been set up, the following experiments
 will be carried out:
\end_layout

\begin_deeper
\begin_layout Itemize
Determining if the efficiency of a wireless mesh network will be severely
 impacted by traffic generated from the underlying management architecture.
\end_layout

\begin_layout Itemize
Determine whether using an external channel for gathering management data
 will significantly improve the situation.
\end_layout

\begin_layout Itemize
Estimate whether a decentralized management architecture is significantly
 better than the conventional method.
\end_layout

\end_deeper
\end_deeper
\end_deeper
\begin_layout Paragraph
(932 words)
\end_layout

\begin_layout Paragraph
A copy of this submission was emailed to my supervisor on 27/03/2007 at
 1pm.
\end_layout

\begin_layout Standard
\begin_inset LatexCommand \bibtex[ieeetr]{Thesis}

\end_inset


\end_layout

\end_body
\end_document