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.

Summer 2006/2007 Internships at CAIA


*** Summer 2006/2007 Internship application date has now passed ***

CAIA is offering five internships over summer 2006/2007 for students who have completed at least 2 years of a Telecommunications Engineering, Electronic Engineering or Computer Engineering degree in Australia. The ideal candidates will have a strong interest in data networking and be at least intensely curious about research in Internet protocol design, development and evaluation.

The internships will be 10 weeks long in total, consisting of three weeks from Monday December 4th 2006 to Friday December 22nd 2006, and seven weeks from Tuesday January 2nd 2007 to Friday February 16th 2007. Applicants will be expected to be available for the entire period.

Potential applicants must make initial contact with CAIA by October 30th 2006.

Successful applicants will work under the supervision of the academic staff member associated with the listed projects.

Projects:

Prospective interns must indicate which of the projects listed below they wish to pursue. Successful applicants will be assigned to a project based on our evaluation of applicant skillset and academic record.

Goal:

Internships are an opportunity to explore life in a research centre, collaborate with experienced academic staff and do a research project outside the boundaries of a regular undergraduate curriculum.  You will begin to appreciate the combination of discipline and imagination that drives modern data networking research.

Stipend:

A tax-exempt stipend of $350/week will be paid to each intern under this scheme.

This internship is limited to students who meet the following criteria:

The first application round is in four parts. (Yes, this is a modest screening process.)

Note: Applications MUST be in pdf (with .pdf extension) or ASCII text (with .txt extension). Applications submitted in MS Word, OpenOffice, or any other format (including snail mail) will simply be ignored.

Final selection of applicants:

We may choose to use followup phone or in-person interviews to make our final selection. More details will be available after November 8th 2006. Where an objective tie-break is required, we will chose the person whose SSH key arrived first (in step 2 above).

Available Projects:

The following projects are available with the nominated academic staff member.


Project 1: Wireless VoIP

Supervisor: Dr Hai Vu
-----------
Voice over IP (VoIP) is an important application in the current and future wireless networks. The aim in this project is to set up a wireless test bed where voice calls can be initiated, as well as traffic measurement can be made in order to understand and improve voice quality and network efficiency.

In particular the wireless test bed will be configured in infrastructure mode which consists of a desktop PC acting as an access point (AP). Mobile stations (clients) are laptop or desktop PC equipped with wireless card. Furthermore, an open source Private Branch eXchange (PBX) called Asterisk will be installed in AP to enable voice calls. This project also involves measuring some characteristic of VoIP in wireless environment such as collision probability, delay and packet loss rate.

Required Skills
---------------
Understanding of the concepts of wireless LAN networks, MAC (802.11) and VoIP (H.324, SIP) protocols. Require good skills in administering FreeBSD and/or other Unix based systems.



Project 2: FIRST PERSON SHOOTER GAME TRAFFIC ANALYSIS


Supervisor: Dr Philip Branch
----------
This project will involve statistical analysis of game traffic captured during trials conducted earlier this year. Within CAIA in the past year we have had considerable success in developing useful and scalable statistical traffic models of game traffic. One of the building blocks in developing traffic models of games with many players is the two-player game. This project will involve the statistical analysis of two-player game traffic generated by a server and transmitted to each player. We have conjectured that the state of the two player game (server to client) can be described in terms of client to server traffic. Some preliminary work analysing Quake 3 seems to support this conjecture but additional statistical analysis is needed to determine whether it is true or false and under what circumstances it is true or false.

Depending on time available further topics in analysing game traffic may be undertaken.

Required skills
---------------
The skills needed to carry out this research project are a good understanding of statistical traffic models (a good mark for HET336 will be highly regarded) and ability in using MATLAB or SCILAB to analyse large datasets.


Project 3: EXPLORING THE PERFORMANCE OF BITTORRENT IN A BROADBAND CONSUMER CONTEXT

Supervisor: Dr Jason But
-----------
Prior research has been done in measuring the performance, efficiency and fairness of BitTorrent as a tool to download and share large files over the Internet [1, 2].  Most of these efforts are directed towards simulation and with networks where all users are equal participants.  In a typical home deployment this is unlikely to be the case as different users are connected to the Internet at different rates.  The aim of this project is to build a virtual network of BitTorrent clients and measure the performance of BitTorrent where different consumer access rates are applied.  We would also like to quantify the interaction between BitTorrent and "other" network traffic sharing the same link and to evaluate the impact on the performance of user-interactive applications of the background traffic generated by BitTorrent.

The project will involve the use of FreeBSD and in particular the configuration of Jails under FreeBSD to implement virtual hosts to run the BitTorrent clients.  The student will install the ctorrent client and configure multiple hosts to perform any required experiments. 

References:
----------
[1] Dongyu Qiu, R. Srikant, "Modeling and performance analysis of BitTorrent-like peer-to-peer networks", ACM SIGCOMM Computer Communication Review , Volume 34 Issue 4, August 2004   

[2] Ashwin R. Bharambe, Cormac Herley, Venkata N. Padmanabhan, "Some observations on bittorrent performance", Proceedings of the 2005 ACM SIGMETRICS international conference on Measurement and modeling of computer systems SIGMETRICS '05,  Volume 33 Issue 1


Required Skills
---------------
A successful applicant would need the necessary skills in administering FreeBSD and/or other Unix based systems with a good understanding of the concepts of process control and management and chroot'ing.


Project 4: Quantifying the enterprise resource costs of dealing with email spam

Supervisors: Associate Professor Grenville Armitage / Associate Professor Richard Constantine (Director, ITS)
-----------

Every day thousands of email spam messages are received, stored, identified, filtered and rejected by Swinburne's corporate email system. This same process occurs at thousands of business around Australia every day. The reception of inbound spam emails incurs network traffic (the TCP connection over which the SMTP message is carried) and CPU resources (processing, storing and checking the inbound email for spam- like characteristics). Because automated spam tagging is not perfect (non-spam may be tagged as spam - known as 'false positives') we also incur a storage cost - emails tagged as spam must be kept for human review, lest we lose important emails that were misclassified as spam. There are finite real-world costs associated with each of these resources, costs that must be borne by the corporate email system.

This project will involve measurement of network traffic associated with the deluge of inbound email received by Swinburne every day, and development of models to calculate the resource consumption attributable to inbound spam traffic. You will work with ITS staff to discover the patterns of resource consumption at the hourly, daily and weekly level associated with spam arrival. Based on your newly acquired knowledge, you will estimate the financial benefits (if any) of spam- mitigation techniques such as post- arrival filtering, DNS black-lists, and network-layer pro-active connection restriction schemes. Along the way you will learn about the need for anonymisation of real-world data and maintaining end-user privacy while developing system usage models.


Required Skills
---------------
Understand how to capture and read IP packet traces with either Ethereal or tcpdump, be capable of installing FreeBSD or Linux onto a regular PC (you'll be using FreeBSD), have a basic understanding of IP networking principles (IPv4 addressing, port numbers, the basic differences between TCP and UDP), understand that 'dir' is not a unix command. Have some basic familiarity with using Excel (or similar tool) to construct graphs or charts from sets of raw data.


Project 5: Minimally Intrusive Round Trip Time Measurements Using Synthetic Packet-Pairs

Supervisors: Mr Sebastian Zander / Associate Professor Grenville Armitage
-----------

In this project we will evaluate a passive technique for round trip time (RTT) estimation called Synthetic Packet-Pairs (SPP). Regular and frequent measurement of round trip time (RTT) between points on the Internet is becoming increasingly important for a range of highly interactive real-time applications (such as multiplayer online games or Voice over IP). Active probing techniques are possible but problematic. The extra packet traffic imposed by active probes along a network path can modify the behaviour of the network under test. In addition, estimated RTT results may be misleading if the network handles active probe packets differently to regular IP packets. We recently published SPP in CAIA Technical Report 060707A, a novel, new approach to passively estimating RTT using IP traffic already in the network. Unlike many existing schemes, SPP works with applications that do not exhibit symmetric client-server packet exchanges (such as many online multiplayer games) and applications generating IP multicast traffic.

Your tasks will involve implementing SPP (in C, C++ or Java) and running a variety of network trials to show how effective SPP can be under different network scenarios. These will include cases where RTT (and finely-grained RTT variation) is measured over load-sensitive link technologies such as wireless LANs, ADSL links and Cable Modems. You will also implement and test improvements to the basic algorithm described in CAIA Technical Report 060707A.

Required Skills
---------------
Understand how to capture and read IP packet traces with either Ethereal or tcpdump, be capable of installing FreeBSD or Linux onto a regular PC (you'll be using FreeBSD), have a basic understanding of IP networking principles (IPv4 addressing, port numbers, the basic differences between TCP and UDP), understand that 'dir' is not a unix command. Have some basic coding skills in C, C++ or Java.

Last Updated: Monday 6-Nov-2006 09:26:52 AEDT | Maintained by: Grenville Armitage (garmitage@swin.edu.au) | Authorised by: Grenville Armitage (garmitage@swin.edu.au)