High Bandwidth Applications and Tools (BoF)
Chairman : Dr. Markus Buchhorn
Members :
Objectives : To identify emerging demands for increasing bandwidth capacity and associated infrastructure across the Asia-Pacific, and tools that enable users to take advantage of new high-speed links.
Target Audience : Network engineers, researchers and other end-users, infrastructure managers
Expected Number of Participants : 30 - 60
Agenda :
  1. Michelle Munson, CEO Aspera/IBM

    Large Data Transport for the Science DMZ - Slides

    One of the most fundamental challenges impeding the successful adoption of high-speed networking and storage in support of today's "big data science" are the challenges of transport large form files and data sets over the WAN. Aspera provides a new approach that allows for moving data between collaborating researchers at global distances with maximum speed. This talk will overview the newest capabilities including direct transfer to/from cloud object storage, bulk data ingest, high-performance file synchronization, and person-to-person collaboration, and will contrast the technology with other forms of data transfer acceleration (GridFTP, UDP data blasters, etc.) It will overview several deployments on public research networks in Europe, the US and Australia by bodies such as NIH, EBI, BGI, and RDSI.

  2. Dr Marek T. Michalewicz (Senior Director of A*STAR Computational Resource Centre, Singapore) and Professor Deng Yuefan (Professor of Stony Brook University, USA; Mt. Taishan Distinguished Professor of National Supercomputer Centre in Jinan, China)

    Inter-campus InfiniBand testbed for RDMA between HPC clusters - Slides

    Supercomputer interconnects are typically InfiniBand links within the same data centre in the same building. Only few cases exist of InfiniBand implementation for data centres across two locations between 1 to ~20 km apart. Here we report our testing of supercomputer interconnects using InfiniBand point-to-point link between multiple campuses over our national R&E networks. The findings of this testbed will lay the foundation for evaluating the possibility of interconnecting supercomputers across very long distances, between countries on different continents.

  3. Yves Poppe (Consultant of A*STAR Computational Resource Centre, Singapore)

    Asia Connects America 100Gbps: the success factors for the ACA-100 project - Slides

    It has been proposed to establish an ACA-100 project, modeled on the successful trans-Atlantic ANA-100 project set up in 2013, which successfully interconnected two 100 Gbps networks, the Internet2 and the GEANT networks in America and Europe respectively. This paper evaluates the current networking situation in Asia, and highlights a few applications, which could be game-changers for advanced research networking.

Seating Arrangement : Classroom
Video Conferencing Facility : Desirable, just in case
Remarks :