Using MapReduce:  MapReduce is a  a software framework introduced by Google to support distributed computing on large data  sets on clusters of computers.  The framework is inspired by map and reducte functions commonly used in functional programming (Wikipedia)..   MapReduce is heavily used by Google.   One of the goals of the Hadoop project is to provide an open source implementation of MapReduce that can run on a clluster or on rented hardware, e.g., Amazon EC2 cluster.  A project could involve the development of a parallel application using Hadoop and Amazon EC2 cluster.  Examples include the following:

• Provide useful information about a set of documents.  For example you can use MapReduce to analyze a set of log files.  An example of this can be found here.   Other sets of documents include blogs and news articles.  
• Develop and run a scientific application on Amazon EC2 cloud.  Describe your experience.     

Implementation of a Subset of MapReduce: MapReduce uses a master process and multiple worker processes for parallel processing. The master process provides work to the work processes.  A project could invovle this aspect of MapReduce with support for electing a new master process in the case that the master process goes down.  More information on implementation issues related to MapReduce can be found here.  

Monitoring Virtual Machines: Dynamic replication is where the number of servers allocated to an application dynamically changes.  This requires suitable monitoring techniques.   The use of virtual machines, however, complicates monitoring since the host OS only sees to VM but not necessarily the processes running in the VM.  Approaches for dealing with this are found in the paper "Black-box and Gray-box Strategies for Virtual Machine Migration".   A possible project could be based on an implementation of one or more of the monitoring techniques described in this paper.

 New Computing Environment for a Department of Computer Science: In this project, you would design  a system (and partially implement) that enhances the computing environment for education purposes.  For example, different courses have different software needs.  Virtual machines can be specifically designed for a course.  Based on this, can we radically change the computing environment found in most computer science departments (at least the one at UWO).