[Storage-research-list] Workshop on Computing with Massive and Persistent Data

Maya Gokhale maya at llnl.gov
Thu Apr 17 12:31:06 EDT 2008


Note: 4-page abstracts due May 30, 2008 (in 6 weeks)

1st Workshop on Computing with Massive and Persistent Data (CMPD'08)



Sponsored by the IEEE Computer Society Mass Storage Systems Technical  
Committee (MSSTC)

September 22, Baltimore, MD, USA
Workshop on Computing with Massive and Persistent Data

Colocated with the Twenty-Fifth IEEE Conference on Mass Storage  
Systems and Technologies
Workshop Description

For many applications, the distinction between primary and secondary  
storage will begin to fade over the next several years. Relatively  
slow improvements in memory latency and bandwidth have already pushed  
the cost of main memory accesses to hundreds of processor cycles.  
Meanwhile, large and inexpensive nonvolatile random access memory  
(NVRAM) devices are defining a new class of storage whose cost,  
capacity, and access time all fall between DRAM and magnetic disk  
drives. Moreover, the current growth in the number of cores per  
processor is not likely to be matched by increases in either main  
memory size or access speed. As a result, processing power is  
becoming cheap compared to memory size, causing applications that  
were once considered compute-intensive to become data-intensive. New  
approaches to file systems and programming models for data-intensive  
applications are also changing many code architects' view of the  
standard hierarchical memory model.

Applications that manipulate massive data structures will be among  
the first to adopt new hardware and software storage technologies.  
Examples include data mining and the analysis of large images and  
graphs. Many of these applications currently use out-of-core  
programming techniques or standard database management systems, which  
both rely on specialized algorithms for working with data that  
resides mostly on disk. Developments in hardware and software  
technology call for a fresh look at these techniques.

This workshop will gather developers and users of new technologies  
for computing with massive and persistent (i.e., nonvolatile) data.  
Areas of interest include:

     o Disk-scale nonvolatile memory devices

     o Other new hardware approaches for computing with massive data  
sets

     o File systems and other system software for massive data sets

     o Novel programming models for data-intensive applications

     o Applications using massive and persistent data

A primary goal of the workshop is to assemble a broad range of  
researchers and practitioners in this field to establish a community  
that crosses traditional boundaries between hardware, system  
software, and applications. We seek to develop a shared view of the  
current state of the art, and of the opportunities and challenges  
that may arise over the next several years. Ultimately, we hope that  
this workshop series can help define research and funding strategies  
in this emerging field.

How to Participate

Prospective participants should submit an abstract of their current  
work or a position paper, not to exceed four pages. Participants will  
be invited attend based on these submissions, though not all  
participants will make a presentation at the workshop. The program  
committee will choose presentations to reflect a diverse set of  
approaches and interests spanning the topics listed above. All  
abstracts and presentations from those chosen to participate will be  
published in the workshop proceedings.

Dates

Position Paper Submission: May 30, 2008. Papers should be no more  
than four pages, at least 10-point font, in PDF or MS-Word Format.  
Email submissions to cmpd08-workshop at llnl.gov

Notification of Acceptance: June 30, 2008.

Final Presentations Due: August 29, 2008. (Presentation slides, in  
PDF or MS-PowerPoint format, must be submitted in advance to vet for  
excessive commercial content, etc., and inclusion in the workshop  
proceedings.)

Workshop Program Committee

Randal Burns, Johns Hopkins University
Maya Gokhale, Lawrence Livermore National Laboratory
Steve Louis, Lawrence Livermore National Laboratory
John May, Lawrence Livermore National Laboratory
Henry Newman, Instrumental, Inc.
Winfried Wilcke, IBM Research


------------------
Maya Gokhale
maya at llnl.gov
925-422-9864


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.andrew.cmu.edu/pipermail/storage-research-list/attachments/20080417/52250a20/attachment.html>


More information about the Storage-research-list mailing list