The following students are examining efficient autonomic strategies for managed (garbage collected) and unmanaged memory on the Android platform. Our "Snappy" version of Google's Android system is an early result of this work.
Alternative Approaches for Low Power computing
We are looking at alternative approaches for use in low power computing, including HPC and mobile devices. We are investigating the application of various ARM architectures to well know HPC applications. This involves characterizing the performance characteristics of various applications and the interaction of the applications with the architectures under investigation. Our work also investigates the possibility of throttling or turning off various system components that are not in use during a computation to save energy.
Bandwidth allocation among elastic and inelastic flows is being investigated by Roger Ochoa.
Manuel Corona's MS thesis work is investigating techniques for reducing the time required to rebuild a RAID system after partial failure.