Odysseus

 
Background:
By 2013, mobile phones and other browser-enabled mobile devices will overtake PCs as the most common web access device worldwide [1]. Every year, the functionality and computing power of the mobile devices have been increasing, with the number of transistors on a chip following the Moore’s law. However, at the same time, the average annual gain in the battery has been improving by only 6% [2]. This has a serious impact on the practical use of the mobile device, especially for multimedia streaming and video-on-demand. For example, the battery of an iPhone 3GS lasts a mere 5 hours during Internet connectivity on a 3G network. Imagine, if a user were to try streaming live video from an iPhone to the internet using the UStream Broadcaster App., the battery would only last for a couple of hours. For a device that must also operate as a phone, these battery deficiencies are unacceptable. Research highlights that three of the biggest drains on the battery of mobile wireless devices are the Wireless Network Interface Card (W-NIC), the screen backlight and the processor.
 
[1] “Gartner's Top Predictions for IT Organizations and Users, 2010 and Beyond: A New Balance”, http://www.gartner.com/resId=1268513, 29 December 2009
[2]            http://www.batteryuniversity.com/parttwo-54.htm
 
Project Idea:
The first step in optimising energy consumption on the above components is to simulate a mobile device in action and analyse its responses to parameter tweaks. Omnet++ is a modular simulation platform written in C++ which is highly customizable through the addition of freely available models. The goal of this project is to combine aspects of a couple of separate models together to provide an intuitive add-on for Omnet++ with functionality for different packetization protocols (UDP, TCP, RTP, DCCP, etc.), mobility patterns (Point-to-point, random, etc.) and battery consumption measurements. This model will then be used to compare different energy-aware solutions for content delivery, including adaptive multimedia streaming.
 
Outcome:
A simulation model and test bed for energy-aware content delivery over wireless and results of comparison-based testing.