Workshop Chairs

• Peter Christen (The Australian National University, Canberra, Australia)
• Bettina Kemme (McGill, Montreal, Canada)
• Erhard Rahm (Univ. of Leipzig, Germany)

Workshop Proceedings

The electronic proceedings of this PhD workshop are available as Volume 1882of the CEUR Workshop Proceedings at http://ceur-ws.org/Vol-1882/.
The entire proceedings in as singe pdf file is available under https://dbs.uni-leipzig.de/file/Proc-vldb-PhD-Workshop2017.pdf.

Preliminary program

• 8:30-10:10 Matching and data integration (chair: Erhard Rahm)

  • Generalizing Matching Knowledge using Active Learning.
    Anna Primpeli, University of Mannheim, Germany
  • Comparing entities in RDF graphs.
    Alina Petrova, Univ. of Oxford, UK
  • Scalable Linkage across Location Enhanced Services.
    Fuat Basik, Bilkent Univ., Turkey
  • Distributed Similarity Joins on Big Textual Data: Toward a Robust Cost-Based Framework.
    Fabian Fier, Humboldt University Berlin, Germany
  
  

• 10:10-10:35 Coffee break

• 10:35-12:15 Data analytics and queries (chair: Timos Sellis)

  • Graph Pattern Mining for Business Decision Support
    André Petermann, Univ. of Leipzig, Germany
  • Efficient Clustering for Large-Scale, Sparse, Discrete Data with Low Fundamental Resolution.
    Veronika Strnadova-Neeley, UC Santa Barbara, USA
  • Query Processing Based on Compressed Intermediates.
    Patrick Damme, Techn. Univ. Dresden, Germany
  • A Hardware-Oblivious Optimizer for Data Stream Processing.
    Constantin Pohl, Techn. Univ. Ilmenau, Germany
  
  

• 12:15-13:30 Lunch break

• 13:30-15:15 Keynote session (chair: Bettina Kemme)

  • Keynote by Prof. Dr. Anastasia Ailamaki, EPFL: The Next 700 Transaction Processing Engines

    Abstract: For over four decades, throughput has been the target metric of choice for Online Transaction Processing engines. Around mid 2000s, however, Dennard scaling came to a crushing halt and now multicore processors provide explicit thread-level parallelism as an alternative to frequency scaling for increasing throughput. Thus, OLTP research focuses on developing scalable synchronization techniques for exploiting parallelism provided by multicore processors. In the late 2000s, DRAM price free-fall made it possible to fit a single server with Terabytes of memory, and to fit most operational databases, with the exception of a few rare cases, entirely in memory. This led to a flurry of research on the design of scalable main-memory OLTP engines that adopt radically different designs compared to their disk-based counterparts. Today, state-of-the-art main-memory OLTP engines can handle millions of transactions per second and provide near-linear scalability under most workloads. However, three recent trends indicate an impending change in OLTP engine design once again: 1) changes in application workloads, 2) shifting hardware landscape, and 3) new target metrics. In this talk, we will discuss the implications of these trends on the design of next-generation transactional engines, and explore new designs with the twin goal of meeting changing application demands and optimizing for the new metrics by exploiting emerging hardware.

    
  • Efficient Migration of Very Large Distributed State for Scalable Stream Processing.
    Bonaventura Del Monte, DFKI Berlin, Germany
  • Spatio-Temporal Locality in Hash Tables.
    Matthew A Pugh, University of Edingburgh, UK
  
  

• 15:15-15:30 Coffee break

• 15:30-16:30 Short presentations (chair: Erhard Rahm)

  • Facilitating User Interaction With Data.
    Zainab Zolaktaf, Univ. of British Columbia Vancouver, Canada
  • Processing Moving Object Data Streams with Data Stream Management Systems
    Tobias L Brandt, University of Oldenburg, Germany
  • Symmetric and Asymmetric Aggregate Function in Massively Parallel Computing
    Chao Zhang, Blaise Pascale University Aubière, France
  • Practical Verification of Hierarchical Artifact Systems.
    Yuliang Li, UC San Diego, USA
  
  
• 16:30-17:30 Panel discussion on PhD research and opportunities (chair: Bettina Kemme)

  Panelists: Mike Carey, Wolfgang Lehner, Themis Palpanas