Lei Xu is a fifth year PhD candidate in the Department of Computer Science. He works on the performance, scalability and reliability of large-scale storage systems, meta-date management, file system design, and manycore operating systems. His publications include “Propeller: A Scalable Metadata Organization for a Versatile Searchable File System” and “SHHC: A Scalable Hybrid Has Cluster for Cloud Backup Services in Data Centers”. In the following, he discusses his work in the Abacus Distribute Storage Lab and how it relates to his dissertation.
What type of work do you do in the Abacus Distribute Storage Lab?
My work in the Abacus Distribute Storage Lab concerns improving the scalability of file systems from difference aspects. First, I am working on designing and implementing a large-scale searchable file system. We believe that it will be challenging to effectively maintain and utilize the large amount of data in the file system, especially for big data applications. Despite this, it has relatively mature technologies to store enormous amounts of data. Finding the interesting data out of the whole dataset is usually complex and time consuming. Therefore, we propose a “searchable” file system, featuring real-time file search capability, to let users retrieve the desired data accurately and efficiently without scanning the file system.
Second, I am currently working on improving the Linux Virtual File System (VFS) scalability on manycore platform. We have found that the current operating system designs are not scalable on the future multicore and manycore platforms. We are taking Linux VFS as a case to investigate the scalability bottlenecks, and propose a distributed VFS design to allow VFS to fit the distributed nature of the NUMA architecture used by the multicore/manycore systems.
How does the work in the lab relate to your studies?
It is the same thing that I am doing my dissertation on; it connects all of the work that I have done over the past four years.
How did you originally get involved with the lab?
I had a storage background and looked for an advisor with a similar background. I introduced myself to Hong Jiang and then joined the lab.
Students who develop an interest in particular areas are encourage to speak with a relevant member of faculty to discuss exploring future research opportunities. Faculty research interests can be found on their webpages, which are listed at http://go.unl.edu/0nn.