THE WORKSHOP ON PERFORMANCE AND SCALABILITY OF STORAGE SYSTEMS (WOPSSS)
The third workshop was held jointly with ISC in Frankfurt, Germany. June 28th, 2018
STATE-OF-THE-ART RESEARCH & INNOVATION, COMING TOGETHER
The Workshop On Performance and Scalability of Storage Systems (WOPSSS) aims to present state-of-the-art research, innovative ideas, and experience that focus on the design and implementation of HPC storage systems in both academic and industrial worlds, with a special interest on their performance analysis.
The arrival of new storage technologies and scales unseen in previous practice lead to significant loss of performance predictability. This will leave storage system designers, application developers and the storage community at large in the difficult situation of not being able to precisely detect bottlenecks, evaluate the room for improvement, or estimate the matching of applications with a given storage architecture. WOPSSS intends to encourage discussion of these issues through submissions of researchers or practitioners from both academic and industrial worlds.
The workshop is held in conjunction with the ISC-HPC during the ISC workshop day. Note that the attendance of ISC workshops requires a workshop pass.
WOPSSS is embedded into a full-day program for I/O that we organize with the HPC-IODC: HPC I/O in the Data Center Workshop team. We will build a program with a joint morning session with keynote talks and best papers and then diverge into two independent workshops in the afternoon. You can learn more at last year's workshop web page.
April 20 May 4
Author Notification: May 31
Camera Ready: June 15
Workshop June: 28
Frankfurt Marriott Hotel
Hamburger Allee 2
60486 Frankfurt am Main
Phil Carns (Argonne National Laboratory)
Understanding Metadata Latency with MDWorkbench
Julian Kunkel and George Markomanolis
Self-Optimized Strategy for IO Accelerator Parametrization
Lionel Vincent and Gaël Goret
Julian M. Kunkel
Performance study of non-volatile memories on a high-end supercomputer
Leonardo Bautista Gomez, Kai Keller and Osman Unsal
Exploring Scientific Application Performance Using Large Scale Object Storage
Steven Wei-der Chien, Stefano Markidis, Rami Karim, Erwin Laure and Sai Narasimhamurthy
From Application to Disk: Tracing I/O Through the Big Data Stack
R. Schmidtke, F. Schintke, and T. Schütt
IOscope: A Flexible I/O Tracer for Workloads’ I/O Pattern Characterization
Abdulqawi Saif, Lucas Nussbaum, and Ye-Qiong Song
Benefit of DDN’s IME-Fuse and IME-Lustre file systems for I/O intensive HPC applications
Eugen Betke and Julian Kunkel
The steering committee has a French bias, since this workshop took its roots as the extension of a similar initiative already running in France.
PROGRAM COMMITTEE (TO BE COMPLETED)
Julien Bigot, CEA, France
Jason Chun Xue, City University of Hong Kong, Honk Kong
Stefano Cozzini, CNR, Italy
Jesus Escudero-Sahuquillo, University of Castilla-La Mancha, Spain
Maria E. Gomez,Polytechnic University of Valencia, Spain
Denis Gutfreund, ATOS, France
Julian Kunkel; DKRZ, Germany
Florin Isaila, University Carlos III, Spain
Duo Liu, Chongqing University, China
Manolis Marazakis, Forth, Greece
Lars Nagel, Johannes Loughborough University, UK
Ramon Nou, BSC, Spain
Juan Piernas Cánovas, Universidad de Murcia, Spain
Josef Weidendorfer, TUM, Germany
Soraya Zertal, University of Versailles, France
POWERED BY THE HPC STORAGE LEADER
DataDirect Networks (DDN) is the world’s leading big data storage supplier to data-intensive, global organizations. For more than 15 years, DDN has designed, developed, deployed and optimized systems, software and solutions that enable enterprises, service providers, universities and government agencies to generate more value and to accelerate time to insight from their data and information, on premise and in the Cloud. Organizations leverage the power of DDN technology and the deep technical expertise of its team to capture, store, process, analyze, collaborate and distribute data, information and content at largest scale in the most efficient, reliable and cost effective manner. DDN customers include many of the world’s leading financial services firms and banks, healthcare and life science organizations, manufacturing and energy companies, government and research facilities, and web and cloud service providers.
Non-Volatile Memory for Next Generation I/O
Dr Michele Weiland, EPCC - University of Edinburgh
The H2020 funded NEXTGenIO project is developing a prototype computing platform that uses on-node non-volatile memory to bridge the latency gap between DRAM and disk, and dramatically improve I/O performance for HPC and data intensive applications. In addition to the hardware that will be built as part of the project, NEXTGenIO is developing the software stack (from OS and runtime support to programming models and tools) that goes hand-in-hand with this new hardware architecture. This talk will look at how the new technology could be used to benefit a wide range of use cases.
The Role of Parallel File Systems in the Era of Burst Buffer and Exascale
Gabriele Paciucci, Scalable Datacenter Solutions group, Intel
New technologies are becoming mainstream to address the I/O requirements of the next generation of HPC clusters. Is the burst buffer the solution for a TB/sec I/O requirement or for, finally, solve the checkpoint use case? Are the upcoming NVRAM changing the game and the design of the next generation of clusters? Which will be the role of the “legacy” parallel file systems? In this presentation we will try to answer those questions and discuss the vision of Intel for the future of the HPC storage.
- Michele Weilland, EPCC, Non-volatile memory for next generation I/O
- Gabriele Paciucci, INTEL, The role of the parallel file systems in the era of burst buffer and Exascale
- George Markomanolis, KAUST, Scientific Applications Performance Evaluation on Burst Buffer
- Julian Kunkel, DKRZ, MPI-IO In-Memory Storage with the Kove XPD
File System Session
- Michael Kuhn, University of Hamburg, JULEA: A Flexible Storage Framework for HPC
- Georgios Bitzes, CERN, Scaling the EOS namespace
- Jakob Blomer, CERN, Delivering LHC software to HPC computing elements with CernVM-FS
- Georgios Koloventzos, BSC, HetFS: A Heterogeneous File System for Everyone
- Discussions and Panels on file system trends