Lab Home | Phone | Search
Center for Nonlinear Studies  Center for Nonlinear Studies
 Home 
 People 
 Current 
 Affiliates 
 Visitors 
 Students 
 Research 
 ICAM-LANL 
 Publications 
 Conferences 
 Workshops 
 Sponsorship 
 Talks 
 Colloquia 
 Colloquia Archive 
 Seminars 
 Postdoc Seminars Archive 
 Quantum Lunch 
 Quantum Lunch Archive 
 CMS Colloquia 
 Q-Mat Seminars 
 Q-Mat Seminars Archive 
 P/T Colloquia 
 Archive 
 Kac Lectures 
 Kac Fellows 
 Dist. Quant. Lecture 
 Ulam Scholar 
 Colloquia 
 
 Jobs 
 Postdocs 
 CNLS Fellowship Application 
 Students 
 Student Program 
 Visitors 
 Description 
 Past Visitors 
 Services 
 General 
 
 History of CNLS 
 
 Maps, Directions 
 CNLS Office 
 T-Division 
 LANL 
 
Thursday, August 10, 2006
3:00 PM - 4:00 PM
CNLS Conference Room

Seminar

Improving File System Usability, Performance and Reliability with Magnetic RAM

Ethan L. Miller
Storage Systems Research Center, University of California, Santa Cruz

Magnetic RAM (MRAM), which is now shipping, has the potential to provide non-volatile, byte-addressable memory for file systems. However, current file systems are not designed to take advantage of byte-addressable RAM, but are instead optimized for block- addressed storage. This talk will describe techniques that we have developed to improve file system performance, usability, and reliability using MRAM. I will first describe mechanisms that drastically reduce the amount of MRAM needed to store metadata by taking advantage of its byte addressability. Next, I will discuss mechanisms for providing richer metadata structures using link structures stored in MRAM. These structures are possible using disk, but MRAM is necessary to make them fast enough for regular use. Finally, I will present techniques we have developed to make MRAM-based file systems orders of magnitude more reliable than disk-based file systems by using online consistency checking and aggressive software-based redundancy. By combining these techniques, we can use MRAM to build faster, more reliable file systems that provide users with better ways to find their data.