FOSDEM'16 HPC, Big Data, and Data Science Devroom

Sunday, January 31, 2016, Brussels, Belgium

Overview

Welcome to the 1st edition of the HPC, Big Data and Data Science devroom, co-located with FOSDEM 2016. FOSDEM is an annual conference about free and open source software, attended by over 5000 developers and open-source enthusiasts from all over the world. FOSDEM has hosted an HPC devroom before, but this is the first time that representatives from the HPC and Big Data communities are joining forces to bring both communities together with this devroom.

  • High Performance Computing (HPC) and Big Data are two important approaches to scientific computing. HPC typically deals with smaller, highly structured data sets and huge amounts of computation while Big Data, not surprisingly, deals with gigantic, unstructured data sets and focuses on the I/O bottlenecks. With the Big Data trend unlocking access to an unprecedented amount of data, Data Science has emerged to tackle the problem of creating processes and approaches to extracting knowledge or insights from these data sets. Machine learning and predictive analytics algorithms have joined the family of more traditional HPC algorithms and are pushing the requirements of cluster and data scalability.

  • Free and Open Source communities have been the foundation of the HPC and Big Data communities for some time. In the HPC community, it should be no surprise that 488 of the Top500 supercomputers in the world run Linux. On the Big Data side, the Hadoop ecosystem has had a tremendous amount of Open Source contributions from a wide range of organizations coming together under the Apache Software Foundation.

  • Our goal is to bring the communities together, share expertise, learn how we can benefit from each other’s work and foster further joint research and collaboration. We welcome talks about Free and Open Source solutions to the challenges presented by large scale computing, data management and data analysis.

The devroom will take place on Sunday, January 31st 2016, at ULB (Campus Solbosch), in Brussels, Belgium. Join us to enjoy a full day of talks, demos and interesting discussions on open-source HPC, Big Data and Data Science.

Sounds interesting? Submit your talk proposal below and see you in Brussels!

Topics

Topics of interest include, but are not limited to:

  • Architecture and design of High Performance Computing (HPC) and Big Data systems
  • Architecture and design of Extract, Transform and Load (ETL) and data acquisition pipelines
  • Data security and governance
  • Tools and technologies related to HPC and computational science, for example:
    • Multithreading (OpenMP, etc.)
    • Distributed computing (MPI, etc.)
    • GPGPU computing (OpenCL, OpenACC, etc.)
    • Parallel filesystems and storage
    • Large-scale performance analysis and debugging
  • Computational paradigms for Big Data systems
    • MapReduce engines
    • Streaming engines
    • SQL engines
    • Dataflow engines
  • Emerging hardware trends of large scale clusters
    • Large scale memory pooling
    • High-speed interconnects
    • ARM cluster architecture
  • System administration of HPC and Big Data clusters
  • User support tools
  • Machine learning libraries and tools
  • Scientific software applications, tools and libraries (across all scientific domains)
  • Big Data platforms, extensions to existing systems, libraries, APIs
  • Experiences reports on using Big Data systems, for example:
    • Large-scale deployments
    • Development and configuration issues
    • Tuning and performance tips and lessons learned
  • Interesting Big Data use-cases and applications
  • Comparative analysis of existing systems, evaluation results, performance studies
  • Interdisciplinary HPC/Big Data use-cases, for example:
    • Applications using both HPC and Big Data technologies
    • Integration issues
    • Open research problems on the convergence of HPC and Big Data
    • Running MPI jobs on Big Data clusters and vice-versa
Submission

We invite presenters to submit talk proposals to present high-quality work with sufficient background material to be clear to the HPC, Big Data, and/or Data Science communities. Talks proposals should be submitted through the FOSDEM Pentabarf server. Submissions must include:

  • Abstract
  • Session type
  • Session length
  • Expected prior knowledge / intended audience
  • Speaker bio
  • Links to code / slides / material for the talk (optional)
  • Links to previous talks by the speaker

Our intention is to have a full day of talks of about 20 minutes each, with an additional 5 minutes for questions by attendees.

We would also like to note:

  • Talks will be streamed live and will be recorded. By submitting a session, speakers agree to being recorded and having their talk made available.
  • All accepted talks will be about (using) free and open source software. We highly discourage “marketing” talks.

When submitting your talk in Pentabarf, make sure to select the ‘HPC, Big Data, and Data Science Devroom’ as the ‘Track’.

If you already have a Pentabarf account from a previous FOSDEM edition, please reuse it. Create an account if, and only if, you don’t have one from a previous year. If you have any issues with Pentabarf, do not despair: contact hpc-bigdata-devroom [at] lists.fosdem.org .

Dates

Call for parcipation available: November 2, 2015

Call for participation closes: December 4, 2015 (there will be no extensions to the deadline)

Devroom schedule available: December 18, 2015

Devroom dinner: Saturday January 30, 2016 (subject to change)

Devroom date: Sunday January 31, 2016

If you would like to create an associated event for the devroom, please fork the page and send a pull request.

Organizers