He..ro DB: A Concept for Parallel Data Processing on Heterogeneous Hardware

Due to the growing demand on processing power and energy efficiency by today’s data-intensive applications developers have to deal with heterogeneous hardware platforms composed of specialized computing resources. These are highly efficient for certain workloads but difficult to handle from the software engineering perspective. Even state-of-the-art database management systems do not exploit all heterogeneous hardware components, as their characteristics differ significantly. They are thus hard to integrate within a coherent database architecture. To address this problem, we propose a design concept that is based on a layered system software architecture: He..ro DB transforms a data-flow graph that describes the data-processing application to a task-based execution plan. Task implementations for the different computing resources and a reasonable degree of parallelism are chosen automatically based on available resources. The concept can cover any hardware configuration and application scenario. It is versatile and offers opportunities for independent optimization on each layer.

[1]  Sebastian Breß.  The Design and Implementation of CoGaDB: A Column-oriented GPU-accelerated DBMS , 2014, Datenbank-Spektrum.

[2]  Gustavo Alonso,et al.  Glacier: a query-to-hardware compiler , 2010, SIGMOD Conference.

[3]  Kunle Olukotun,et al.  Hardware acceleration of database operations , 2014, FPGA.

[4]  Peter Benjamin Volk,et al.  GPU join processing revisited , 2012, DaMoN '12.

[5]  Bharat Sukhwani,et al.  Large Payload Streaming Database Sort and Projection on FPGAs , 2013, 2013 25th International Symposium on Computer Architecture and High Performance Computing.

[6]  Cédric Augonnet,et al.  StarPU: a unified platform for task scheduling on heterogeneous multicore architectures , 2011, Concurr. Comput. Pract. Exp..

[7]  Viktor K. Prasanna,et al.  Fast Regular Expression Matching Using FPGAs , 2001, The 9th Annual IEEE Symposium on Field-Programmable Custom Computing Machines (FCCM'01).

[8]  Sebastian Breß,et al.  Why it is time for a HyPE: A Hybrid Query Processing Engine for Efficient GPU Coprocessing in DBMS , 2013, Proc. VLDB Endow..

[9]  Jon Crowcroft,et al.  Unikernels: library operating systems for the cloud , 2013, ASPLOS '13.

[10]  Gustavo Alonso,et al.  Runtime Parameterizable Regular Expression Operators for Databases , 2016, 2016 IEEE 24th Annual International Symposium on Field-Programmable Custom Computing Machines (FCCM).

[11]  Bingsheng He,et al.  OmniDB: Towards Portable and Efficient Query Processing on Parallel CPU/GPU Architectures , 2013, Proc. VLDB Endow..

[12]  Volker Markl,et al.  Hardware-Oblivious Parallelism for In-Memory Column-Stores , 2013, Proc. VLDB Endow..

[13]  Gustavo Alonso,et al.  Sorting networks on FPGAs , 2012, The VLDB Journal.

[14]  R. Govindarajan,et al.  Fluidic Kernels: Cooperative Execution of OpenCL Programs on Multiple Heterogeneous Devices , 2014, CGO '14.

[15]  Gregory Diamos,et al.  Harmony: an execution model and runtime for heterogeneous many core systems , 2008, HPDC '08.

[16]  Tilmann Rabl,et al.  Generating custom code for efficient query execution on heterogeneous processors , 2017, The VLDB Journal.

[17]  Bharat Sukhwani,et al.  A Hardware/Software Approach for Database Query Acceleration with FPGAs , 2014, International Journal of Parallel Programming.

[18]  Meikel Pöss,et al.  New TPC benchmarks for decision support and web commerce , 2000, SGMD.

[19]  Moriyoshi Ohara,et al.  A Dynamically Reconfigurable Equi-Joiner on FPGA , 2015 .