[HPDC-announce] 4th CFP: 1st Programmable File Systems Workshop (PFSW at HPDC'14)
viji_06 at hotmail.com
Thu Feb 20 01:33:26 CST 2014
Thanks for the information, i would like to share our event link with you.
From: carlosm at soe.ucsc.edu
Date: Tue, 18 Feb 2014 01:49:57 -0800
To: storage-research-list at ece.cmu.edu; distributed-computing-announce at datasys.cs.iit.edu; HPDC-announce at datasys.cs.iit.edu; hpc-announce at mcs.anl.gov
CC: pat at lanl.gov
Subject: [HPDC-announce] 4th CFP: 1st Programmable File Systems Workshop (PFSW at HPDC'14)
[Apologies if you have received multiple copies of this email]
1st Programmable File Systems Workshop (PFSW’14)
in conjunction withThe 23rd International ACM Symposium on High Performance Parallel and Distributed Computing (HPDC 2014)Vancouver, BC, Canada on June 23-27, 2014 (workshop is one day TBD)
WORKSHOP ABSTRACTA major milestone in the evolution of digital computers was the development ofthe stored-program concept and the design of Turing-complete machines asopposed to fixed-program computers. Yet, we still treat an increasinglyimportant subsystem of computers largely as a fixed-program computer: file andstorage systems. Among the key reasons for this history is the justified fearthat (1) any interface changes in file and storage systems will make legacydata inaccessible and locks the data to a particular system and (2)programmability will increase the probability of data loss.
Yet with the advent of open source file systems a new usage pattern emerges:users isolate subsystems of these file systems and put them in contexts notforeseen by original designers. Examples are: (1) an object-based storage backend gets a new RESTful front end to become a Amazon Web Service's S3 compliantkey value store, (2) a data placement function is used as a placement functionfor customer accounts, and (3) the HDF5 scientific data access library isembedded into parallel storage systems. This trend shows a desire for theability to use existing file system services and compose them to implement newservices — a desire, however, that is frequently stumped by the difficulty ofbringing new services of advanced functionality up to production quality andsufficiently low probability of data loss. At the same time government andindustry are heavily investing into the development of new, extremelyscalable, and highly efficient, distributed I/O stacks that largely abandontraditional file and storage system interfaces.
Designing programmability into file and storage systems has the followingbenefits: (1) we are achieving greater separation of storage performanceengineering from storage reliability engineering, making it possible tooptimize storage systems in a wide variety of ways without risking years ofinvestments into code hardening; (2) we are creating an environment thatencourages people to create a new stack of storage systems abstractions, bothdomain-specific and across domains, including sophisticated optimizers thatrely on machine learning techniques; (3) we are informing commercial parallelfile system vendors on the design of low-level APIs for their products so thatthey match the versatility of open source storage systems without having torelease their entire code into open source; and (4) we are using thishistorical opportunity to leverage the tension between the versatility of opensource storage systems and the reliability of proprietary systems to lead thecommunity of storage system designers.
GOALThis one-day workshop focusses on frameworks that allow the programmability offile and storage systems while addressing the risks of data interface change.The workshop aims to serve as a venue for leaders in the file system andstorage community to exchange ideas outside the tradition of half a century ofclassic file and storage systems research which focussed on a small set ofunchanging interfaces.
PAPER SUBMISSIONSAuthors are invited to submit papers with unpublished, original work of notmore than 8 pages of double column text using single-spaced 10 point size on8.5 x 11 inch pages (including all text, figures, references, and appendices),as per ACM 8.5 x 11 manuscript guidelines (document templates can be found athttp://www.acm.org/sigs/publications/proceedings-templates). Electronicsubmissions in pdf format are received athttps://www.easychair.org/conferences/?conf=pfsw2014 at the submissiondeadline.
TOPICSAddressing programmability of the non-volatile part of the memory hierarchy,the workshop seeks contributions on relevant topics, included but not limitedto:
- Programming models - Data interface change management and isolation - Interface metadata management and propagation - Compile-time and runtime storage optimization - Data and task placement in large-scale storage stack - Local and distributed performance management and isolation- Nonstop storage system evolution
IMPORTANT DATESSubmission of papers: February 21, 2014, 11:59 PM PSTAuthor notification: March 18, 2014Camera-ready papers: April 15, 2014Workshop: One day during June 23-27, 2014
WORKSHOP ORGANIZERSCarlos Maltzahn - University of California, Santa CruzPatrick McCormick - Los Alamos National Laboratory
PROGRAM COMMITTEEJohn Bent, EMCAndré Brinkmann, Johannes Gutenberg-Universität MainzRandal Burns, Johns Hopkins UniversityPhil Carns, Argonne National LaboratoryYong Chen, Texas Tech UniversityToni Cortes, Universitat Politècnica de Catalunya Evan Felix, Pacific Northwest National LaboratoryMaya Gokhale, Lawrence Livermore National LaboratoryGary Grider, Los Alamos National LaboratoryDean Hildebrand, IBM AlmadenDries Kimpe, Argonne National LaboratoryScott Klasky, Oak Ridge National LaboratoryQuincey Koziol, HDF GroupJay Lofstead, Sandia National LaboratoryBarney Maccabe, Oak Ridge National LaboratoryCarlos Maltzahn, University of California at Santa CruzAdam Manzanares, HGSTPat McCormick, Los Alamos National LaboratoryMichael Mesnier, IntelKiran-Kumar Muniswamy-Reddy, Amazon.comNeoklis Polyzotis, University of California at Santa CruzRob Ross, Argonne National LaboratorySage Weil, Inktank StorageBrent Welch, GoogleJon Woodring, Los Alamos National Laboratory
Associate Adjunct Professor
Computer Science Department
University of California, Santa Cruz
HPDC-announce mailing list
HPDC-announce at datasys.cs.iit.edu
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the HPDC-announce