|
|
|
Scientists usually have to process their data generated from
computer simulations or experiment instruments on their desktop
machines. These data processing tasks are increasingly limited
by the storage capacity and performance provided by a single
workstation, as both data sizes and processing complexity
are growing rapidly. Meanwhile,
a large portion of the storage resources equipped at desktop computers
sit unused. In this project, we investigate how to aggregate the
distributed idle storage resources in a LAN environment, to
create a unified cache/scratch space that holds
hot datasets.
In other words, our research tries to help scientists avoid being forced to use the approach suggested in the following Dilbert strip:
Our research covers many interesting aspects of the storage aggregation problem, with a focus on the following issues:
|
|
Faculty Collaborators
|
