Overview

Building upon the success of the First International Active Workshop at ICDE'17 conference, the second edition of Active will focus primarily on Active Middleware. The objective of this one-day workshop is to investigate opportunities in exploiting the crossroad between novel hardware technologies and middleware: active (compute-intensive) technologies such as active memory, active networking, and active storage for accelerating distributed data-intensive workloads, and investigate issues in enabling these capabilities by revisiting the entire middleware stack hosted on cloud. Our mission is to inspire a creative workshop atmosphere, where everyone contributes to a lively discussion during the entire course of the event.

Scope

Recent popularity of Big Data applications has renewed interest in developing and optimizing distributed data-intensive workloads. In addition to the traditional scientific computing domain, Big Data phenomena are observed in a variety of new domains such as Internet of Things (IoT), social analytic, personalized health and precision medicine, bioinformatics, energy informatics, and emergency response and disaster management. The Big Data workloads are characterized by unprecedented volumes and velocity of data, both historic and transient, very high-speed data flows (e.g., from sensor networks), and diverse variety of data from completely unstructured text documents, to structured relational tables or matrices, to photos and videos. Analyzing vast quantities of such complex data (partially powered by machine learning) is becoming as important as the traditional massive-scale data management.

Unfortunately, existing approaches to solve data-intensive problems are woefully inadequate to address the challenges raised by the Big Data applications. Specifically, these approaches require data to be processed to be moved near the computing resources while studying the role of middleware paradigm to this end. These data movement costs can be prohibitive for large data sets such as those observed in the aforementioned workloads. One way to address this problem is to bring virtualized computing resources closer to data, whether it is at rest or in motion. This is the premise of "active" middleware that enable key data storage and communication components to compute on the persistent or transient data.

Although prototypes of systems with active technologies are currently available, there is a very limited exploitation of their capabilities in real-life problems. The proposed workshop aims to evaluate different aspects of the active middleware stack and understand impact of active technologies on different applications workloads. Specifically, the workshop hopes to understand the role of modern hardware to enable active medium (whether network, storage, memory) over the entire path and the lifecycle of data especially as today's distributed systems opt to hierarchies of storage and memory. Furthermore, we aim to revisit the interplay between algorithmic modeling, compiler and programming languages, virtualized runtime systems and environments, and hardware implementations, for effective exploitation of active technologies.


Program

Tuesday, December 12, 2017

Opening Remarks (8:45-9:00am)

    Session 1: Keynote (9:00-10:00am)

    • Hans-Arno Jacobsen
      "Online Reconfigurable Line-rate Matching of Filter Rules"

    Break (10:00-10:30am)

      Session 2: Research Paper & Invited Student Talks (10:30-12:00pm)

      • Xu Cui, Michael Mior, Bernard Wong, Khuzaima Daudjee, Sajjad Rizvi (University of Waterloo)
        "NetStore: Leveraging Network Optimizations to Improve Distributed Transaction Processing Performance" [Research Paper, Slides]
      • Sukanya Bhowmik (University of Stuttgart)
        "Harnessing the Power of Software-Defined Networking for a High Performance Communication Middleware"
      • Daniel Ritter (University of Vienna)
        "Hardware Accelerated Application Integration: Challenges and Opportunities" [Slides]

      Lunch (12:00-1:30pm)

        Session 3: Invited Talks (1:30-3:00pm)

        • Peter Pietzuch (Imperial College London)
          "SABER: Hybrid Data Processing with Heterogeneous Servers"
        • Timothy Wood (George Washington University)
          "Active Network Middleware with OpenNetVM"

        Break (3:00-3:30pm)

          Session 4: Next Big Problems Panel (3:30-4:30pm)

            Closing Remarks (4:30-4:45pm)

            Topics of Interest

            (not limited to the following)

            • Novel Architectures and Architectural Extensions

              Type of "active" components by exploiting emerging modern hardware such as FPGAs, GPUs, and ASICs
              Active blockchains fabric (shared, replicated and distributed ledgers)
              Active and (weak) consistent distributed caching (hierarchical volatile and non-volatile memory) by exploiting merging hardware such as HTM, RDMA, NVM
              Active distributed replication and high-availability
            • Transaction on Modern Hardware

              Distributed transaction models and the role of ACID
              Transactional memory
              Transactional caching using RDMA
              Transactions in rack- and geo-scale architectures
              Transactions in rack- and geo-scale architectures
            • Distributed Coordination and Consistency Models

              Hardware-assisted agreement protocols (2PC, 3PC, Quorum, Consensus, PAXOS)
              Fast hardware-aware commit protocols (e.g., RDMA)
              Instant hardware-aware recovery, fail-over, replication protocols (e.g., RDMA)
              Hardware-assisted ordering (e.g., group commit protocols)
              Hardware-assisted reliable and persistent network
              Hardware-assisted network and sites fault-tolerance and high-availability
            • Middleware on software-hardware-system co-design

              Co-processor design by offloading computation to accelerators
              Co-placement design by placing accelerator on the path of data (partial computation or best effort computation)
              Co-design virtualization on cloud
              Co-design privacy and security concerns
              Co-design approximate in-network computing
            • Distributed programming paradigms and algoritmic modeling

              Exposing active elements to the programmers---language extensions/directives/pragmas, libraries, and intrinsic
              Accelerator-based programming abstractions (offloading, placement, cooperation)
              Code generations and domain-specific languages (DSLs)
              Novel abstract execution cost models over heterogeneous hardware
              Novel compiler optimization on heterogeneous platforms
              Persistent programming paradigm
              Persistent abstractions for networks
              Middleware abstraction for cross-platform interoperability
              Middleware abstraction to unity heterogeneous hardware
            • Middleware Adaptability

              Adaptive and evolutionary approaches
              Workload-adaptability (scalability and elasticity)
              Hardware-assisted data partitioning, clustering, replication, and distribution
              Energy-adaptability and energy-aware middleware architectures

            Organization


            Workshop Co-Chairs

            Publicity Chair

            • Kaiwen Zhang (Technical University of Munich)

            Program Committee

            • Carsten Binnig (Brown University)
            • Sebastian BreƟ (German Research Center for Artificial Intelligence)
            • Bingsheng He (National University of Singapore)
            • Alessandro Margara (Politecnico di Milano)
            • Tilmann Rabl (TU Berlin)
            • Kai-Uwe Sattler (Ilmenau University of Technology)
            • Jens Teubner (TU Dortmund)
            • Vassilis J. Tsotras (University of California - Riverside)
            • Stratis Viglas (Google Inc., University of Edinburgh)

            Important Dates


            Paper submissions (Extended): September 15, 2017
            Notification to authors: October 6, 2017
            Camera-ready copy due: October 20, 2017
            Workshops: December 11-15, 2017

            Submission Instructions


            Submitted papers must adhere to the formatting instructions of the ACM SIGPLAN style, which can found on the ACM template page. A sample Latex template using the correct ACM SIGPLAN style can be found here. All accepted papers will appear in a Middleware 2017 companion proceedings, which will be available in the ACM Digital Library. Submitted papers can be of three kinds (all accepted papers will be presented at the workshop):

            • Regular Research Papers: These papers should report original research results or significant case studies/experimental surveys.

              They should be at most 8 pages.
            • Position Papers: These papers should report novel research directions or identify challenging problems.

              They should be at most 4 pages.
            • Abstracts: These papers should report on preliminary and ongoing work or exploratory research directions.

              They should be at most 1 page.

            Papers have to be submitted electronically as PDF files via HotCRP.