SimFS: A Simulation Data Virtualizing File System Interface

Nowadays simulations can produce petabytes of data to be stored in parallel filesystems or large-scale databases. This data is accessed over the course of decades often by thousands of analysts and scientists. However, storing these volumes of data for long periods of time is not cost effective and, in some cases, practically impossible. We propose to transparently virtualize the simulation data, relaxing the storage requirements by not storing the full output and re-simulating the missing data on demand. We develop SimFS, a file system interface that exposes a virtualized view of the simulation output to the analysis applications and manages the re-simulations. SimFS monitors the access patterns of the analysis applications in order to (1) decide the data to keep stored for faster accesses and (2) to employ prefetching strategies to reduce the access time of missing data. Virtualizing simulation data allows us to trade storage for computation: this paradigm becomes similar to traditional on-disk analysis (all data is stored) or in situ (no data is stored) according with the storage resources that are assigned to SimFS. Overall, by exploiting the growing computing power and relaxing the storage capacity requirements, SimFS offers a viable path towards exa-scale simulations.

[1]  Joachim Stadel,et al.  PKDGRAV3: beyond trillion particle cosmological simulations for the next era of galaxy surveys , 2016, 1609.08621.

[2]  B. Fryxell,et al.  FLASH: An Adaptive Mesh Hydrodynamics Code for Modeling Astrophysical Thermonuclear Flashes , 2000 .

[3]  Olivier Coulaud,et al.  Toward a Computational Steering Environment for Legacy Coupled Simulations , 2007, Sixth International Symposium on Parallel and Distributed Computing (ISPDC'07).

[4]  Michael E. Papka,et al.  The web page , 2000 .

[5]  Michel Dubois,et al.  Cache replacement algorithms with nonuniform miss costs , 2006, IEEE Transactions on Computers.

[6]  Fan Zhang,et al.  Enabling In-situ Execution of Coupled Scientific Workflow on Multi-core Platform , 2012, 2012 IEEE 26th International Parallel and Distributed Processing Symposium.

[7]  Torsten Hoefler,et al.  Scientific Benchmarking of Parallel Computing Systems Twelve ways to tell the masses when reporting performance results , 2017 .

[8]  Song Jiang,et al.  LIRS: an efficient low inter-reference recency set replacement policy to improve buffer cache performance , 2002, SIGMETRICS '02.

[9]  Mark D. Hill,et al.  21st century computer architecture , 2014, PPoPP '14.

[10]  Peter Braam,et al.  The Lustre Storage Architecture , 2019, ArXiv.

[11]  Joonwon Lee,et al.  CFLRU: a replacement algorithm for flash memory , 2006, CASES '06.

[12]  Ray W. Grout,et al.  Ultrascale Visualization In Situ Visualization for Large-Scale Combustion Simulations , 2010 .

[13]  Torsten Hoefler,et al.  Designing Bit-Reproducible Portable High-Performance Applications , 2014, 2014 IEEE 28th International Parallel and Distributed Processing Symposium.

[14]  Nimrod Megiddo,et al.  ARC: A Self-Tuning, Low Overhead Replacement Cache , 2003, FAST.

[15]  Karsten Schwan,et al.  Flexible IO and integration for scientific codes through the adaptable IO system (ADIOS) , 2008, CLADE '08.

[16]  Torsten Hoefler,et al.  Near-global climate simulation at 1 km resolution: establishing a performance baseline on 4888 GPUs with COSMO 5.0 , 2017 .

[17]  Nicolay J. Hammer,et al.  An online theoretical virtual observatory for hydrodynamical, cosmological simulations , 2016 .

[18]  Fan Zhang,et al.  Combining in-situ and in-transit processing to enable extreme-scale scientific analysis , 2012, 2012 International Conference for High Performance Computing, Networking, Storage and Analysis.

[19]  André Brinkmann,et al.  Analysis of the ECMWF Storage Landscape , 2015, FAST.

[20]  Michel Dubois,et al.  Cost-sensitive cache replacement algorithms , 2003, The Ninth International Symposium on High-Performance Computer Architecture, 2003. HPCA-9 2003. Proceedings..

[22]  L. Sedov Similarity and Dimensional Methods in Mechanics , 1960 .

[23]  Gustavo Alonso,et al.  Reproducible Floating-Point Aggregation in RDBMSs , 2018, 2018 IEEE 34th International Conference on Data Engineering (ICDE).

[24]  Eric Winsberg,et al.  Science in the Age of Computer Simulation , 2010 .

[25]  Alan R. Duffy,et al.  THE THEORETICAL ASTROPHYSICAL OBSERVATORY: CLOUD-BASED MOCK GALAXY CATALOGS , 2014, 1403.5270.