As part of a broader organisational restructure, data networking research at Swinburne University of Technology has moved from the Centre for Advanced Internet Architecture (CAIA) to the Internet For Things (I4T) Research Lab.

Although CAIA no longer exists, this website reflects CAIA's activities and outputs between March 2002 and February 2017, and is being maintained as a service to the broader data networking research community.

Project Ideas

This page lists undergraduate and minor thesis project ideas, suitable for CAIA Internships, research-oriented HET552/HET550/556 projects, and/or MSc(NS) minor theses.

Short projects of a few weeks duration might be suitable for undergraduate projects and internships. Long projects may be suitable for a solid half year research effort (e.g. a minor thesis in a Masters by coursework program). Contact Professor Grenville Armitage,  Associate Professor Hai Vu, Associate Professor Lachlan Andrew, Dr Philip Branch, Dr Jason But or Mr Tony Cricenti if you are interested in pursuing one of them. Alternatively, propose something else along these lines or more generally in one of the centre's research areas.

Completed projects can be found here.


Short Projects
Long Projects

bullet

Fairness of usage caps in Internet access services

bullet

Lawful interception of VoIP survey

bullet

Bit Torrent stealing bandwidth

bullet Transparent wake-on-LAN server
bullet Network aware sleeping of PCs
bullet

A business case for Australia as the world's backup server farm

bullet

Bit Torrent fairness

bullet

Rapid identification of traffic classes

bullet

Internet Lawful Interception system development

bullet

Using network measurement tools to estimate the traffic load between Australia and the rest of the world

Short projects

Fairness of usage caps in Internet access services

Some Australian Internet Service Providers (ISPs) manage customer network use through Usage Caps - a limit on the number of megabytes your transfer each month before incurring extra charges or reductions in available bandwidth. An important question is how many of those (mega)bytes were actually meaningful or useful to the consumer. Everytime a consumer surfs a web page or streams media using http, downloads files using ftp, or sends and receives email that consumer's computer sends packets using the Transmission Control Protocol, or TCP. TCP ensures 'reliable' data transfer for http, ftp and similar applications by retransmitting packets that are lost by the ISP. So if the ISP's network is lossy, and they count your retransmitted packets towards your monthly usage cap, who loses?

For this project you'll be required to take samples of typical IP traffic in and out of a host computer (e.g. using unix/FreeBSD tools such as tcpdump) and derive an estimate of real-world packet retransmission statistics. Evaluate how much data a consumer gets to successfully transfer at the application level (above TCP) before they hit the raw usage caps imposed by their ISP. Develop a commentary on the incentives an ISP might have (or not have) to reduce packet loss rates in their network.

Grenville Armitage. Current until December 2013.

Lawful Interception of the VOIP traffic

One of the key regulatory requirements of any publicly offered communications service is that, subject to judicial oversight, it be able to be intercepted. This project will survey the different approaches to intercepting VoIP traffic and discuss the strengths and weaknesses of different approaches.

Philip Branch. Current until December 2013.

Bit Torrent stealing bandwidth

BitTorrent is a popular protocol used to download large files over the Internet. Unlike traditional client/server models, BitTorrent is actually faster as the number of downloaders increase, as users participate in the download process by uploading portions of the file to other users. This project involves measuring the decrease in available throughput of other (interactive) applications when BitTorrent is sharing a home consumer link. This will also involve measuring the effects of bandwidth limiting the BitTorrent client as well as prioritising non-BitTorrent traffic.

Jason But. Current until December 2013.

Transparent wake-on-LAN server

Current PCs have many energy-saving features, which are seldom used because of inconvenience. For example, most PCs have a "wake on LAN" feature which allows them to be turned on remotely when needed. Still, many people leave computers running "just in case" they need to access them remotely. In this project, a PC will be configured to proxy remote access requests, and transparently wake up the target computer if that computer is asleep. This will require determining whether the target computer is currently asleep, and detecting when it has woken up. This project is likely to require some kernel programming in either FreeBSD or Linux. It will develop skills in the rapidly emerging
"green IT" industry.

Lachlan Andrew. Current until December 2013.

Network aware sleeping of PCs

.Some people leave computers running permanently so that their applications are just the same when the come back the next day. Most computers allow a "hibernate" or "suspend-to-disk" option which has the same effect -- except that it closes existing TCP sessions, which kills such as applications ssh or telnet. In this project, a tool will be developed to warn the user of which connections will be closed when a PC is put to sleep.

Lachlan Andrew. Current until December 2013.
 

Long projects

A business case for Australia as the world's backup server farm

With so much information now being stored digitally, disaster recovery plans require highly robust and reliable backup servers to hold sensitive information. Australia is geographically isolated, politically stable, has a technologically advanced telecommunications infrastructure, and is less than 300milliseconds from most well-connected places on the Internet. Develop a server farm solution that offers secure, offsite backup services to Northern Hemisphere companies and organizations. Identify the major requirements that must be met before such a business could be profitable (e.g. estimated running costs of a secure location, bandwidth in and out of Australia, TCP performance over long paths, etc)  Your solution should support multiple independent customers, scale from gigabytes to terabytes per customer, and provide independent "24 by 7" secure access over the Internet. Make use of existing off-the-shelf TCP/IP based encryption and file transfer/sharing applications and techniques (e.g. open source unixes such as FreeBSD or Linux for file servers, IPsec encryption for tunneling between servers and customer sites, etc)

Grenville Armitage. Current until December 2013.

Bit Torrent fairness

BitTorrent is a popular protocol used to download large files over the Internet. Unlike traditional client/server models, BitTorrent is actually faster as the number of downloaders increase, as users participate in the download process by uploading portions of the file to other users. This project involves a comparison of a number of concurrent BitTorrent clients at the ends of different download pipes (eg. 1.5Mb/s and 512kbps) and to compare the fairness of the download protocol - does the lower bandwidth client get a fair share of download capacity and do all clients fairly contribute to uploads?

Jason But. Current until December 2013.

Rapid identification of traffic classes

We have shown that machine learning can be used very effectively to identify a specific application that generates a traffic flow. However, using machine learning to identify whole classes of traffic is much more challenging. This project will investigate whether a classifier can be constructed that can rapidly identify broad traffic classes such as VoIP, real-time games or video conferencing.

Philip Branch and Jason But. Current until December 2013.

Internet Lawful interception system

This project will further develop a prototype lawful interception system for IP traffic that integrates mechanisms for ensuring that data captured and reported has not been tampered with.

Philip Branch. Current until December 2013.

Using network measurement tools to estimate the traffic load between Australia and the rest of the world

It is of great interest to ISPs to configure and manage their networks in an efficient way and facilitate the efficient use of network bandwidth for Internet traffic. Although the traditional measurement tools such as ping, traceroute and ttcp have been used with great success, they do not measure advanced performance metrics, such as bandwidth capacity, available bandwidth or bottleneck location across a routing path. In this project, we propose the use of the best-current-practice network measurement tools to measure these advanced network performance metrics on the Internet routing paths between Australia and the outside world (i.e., Asia, Europe and America).

Grenville Armitage. Current until December 2013.

Last Updated: Thursday 29-Nov-2012 09:25:42 AEDT | Maintained by: Philip Branch (pbranch@swin.edu.au) | Authorised by: Grenville Armitage ( garmitage@swin.edu.au)