European Extreme Performing Big Data Stacks

Risultati finali

Report on data made available to the ORDP

A report analysing the way each of the E2Data’s components will interact with the standard platforms of choice along with the whole architecture.

Dissemination and exploitation activities report 2

Reports the dissemination activities performed during each year of the project and those planned for the forthcoming period. Additionally, reports the consortium intentions and activities for the successful exploitation/commercialization of project results throughout the project as well as beyond that.

Standardization report

Describes the activities carried out within the project towards industry standardization of the project results including preparation of submissions to standards grouping, consensus building within the standardization communities, and addressing any technical challenges or inclusion of alternative approaches. The report will describe the progress the project has achieved through the standardization process within the appropriate standards bodies.

Prototype V3 and final evaluation

Demonstration of the correct operation of entire E2Data system on large-scale data sets. Demonstration of the correct operation of heterogeneous based accelerators on both x86 and ARM Architectures. Report on performance/power gains on all use cases and achievement of KPIs.

Dissemination and exploitation activities report 1

Reports the dissemination activities performed during each year of the project and those planned for the forthcoming period. Additionally, reports the consortium intentions and activities for the successful exploitation/commercialization of project results throughout the project as well as beyond that.

Cloud Software and Big Data Stack Implementation V2

Regarding Apache Flink, this deliverable will be realized through an external Java Archive (JAR) file with a properly and clean defined API for integration. In addition, documentation in the form of Java Doc will be provided for this package. With respect to OpenStack, this deliverable will be realized through a Linux open source install package that when applied to the target platform will install and make available the Delegator features and API available to the platform. This will include the API definition document in Linux man format.

Intermediate API definition

The intermediate API for both Big Data frameworks and Cloud Providers will be internally used for delivering D3.2.

The Intelligent Scheduler Definition

The architecture of the E2Data Intelligent Scheduler is defined. The definition document will detail the interactions and APIs between Tasks 4.2, 4.3, and 4.4.

Final API definition

The API for both Big Data frameworks and Cloud Providers will be finalized, published and made available to the Delegator installation packages.

User requirements and data management plan

This deliverable will report the requirements of the intended user groups that will drive the definition of E2Data architecture. Finally, the Data Management Plan will be released with any possible data releases, if possible, to the EU Open Data portal.

Prototype V2 and intermediate evaluation

Demonstration of the correct operation of entire E2Data system on large-scale data sets. Demonstration of the correct operation of heterogeneous accelerators on both x86 and ARM Architectures. Report on performance/power gains on available benchmarks and use cases.

Quality plan

This deliverable will describe the quality plan of the project and the required activities for the efficient implementation of the quality system. It will also include a simple but state-of-the-art risk management plan.

Dissemination and exploitation activities report 3

Reports the dissemination activities performed during each year of the project and those planned for the forthcoming period. Additionally, reports the consortium intentions and activities for the successful exploitation/commercialization of project results throughout the project as well as beyond that.

Prototype V1 and preliminary evaluation

Initial prototype installation and standard benchmarking on x86 and ARM (CPU only).

Heterogeneous Execution Prototype V2

Following the continuous requirement analysis of the use cases, all identified kernels will be offloaded and executed on CPU/GPGPU/FPGA/Xeon Phi units for both x86 and ARM architectures.

Profiling tool

Based on deliverables from WP4 and WP5 the profiling compilation infrastructure along with the profiling history of the JVM will be implemented for all targeted architectures.

Visualization tool

This deliverable will be realized as a new E2Data version of Apache Flink’s Web Client.

Project Presentation and web portal

A project presentation and the initial project website with elementary information over the project and its logo.

The Intelligent Scheduler prototype V2

Report and Prototype on the final version of the E2Data Intelligent Scheduler. Includes the final versions of the Intelligence, Decision Making and Scheduling modules of E2Data.

Pubblicazioni

An Overview of Hawk: A Hardware-Tailored Code Generator for the Heterogeneous Many Core Age

Autori: Breß, S., Funke, H., Zeuch, S., Rabl, T. & Markl, V.
Pubblicato in: BTW 2019 – Workshopband., 2019
Editore: Gesellschaft für Informatik
DOI: 10.18420/btw2019-ws-07

Towards Prototyping and Acceleration of Java Programs onto Intel FPGAs

Autori: Michail Papadimitriou, Juan Fumero, Athanasios Stratikopoulos, Christos Kotselidis
Pubblicato in: 2019 IEEE 27th Annual International Symposium on Field-Programmable Custom Computing Machines (FCCM), 2019, Pagina/e 310-310, ISBN 978-1-7281-1131-5
Editore: IEEE
DOI: 10.1109/fccm.2019.00051

Generating Reproducible Out-of-Order Data Streams

Autori: Philipp M. Grulich, Jonas Traub, Sebastian Breß, Asterios Katsifodimos, Volker Markl, Tilmann Rabl
Pubblicato in: Proceedings of the 13th ACM International Conference on Distributed and Event-based Systems - DEBS '19, 2019, Pagina/e 256-257, ISBN 9781-450367943
Editore: ACM Press
DOI: 10.1145/3328905.3332511

A Methodology for Saving Energy in Educational Buildings Using an IoT Infrastructure

Autori: Georgios Mylonas Dimitrios Amaxilatis Stelios Tsampas Lidia Pocero Joakim Gunneriusson
Pubblicato in: 10th International Conference on Information, Intelligence, Systems and Applications (IISA 2019), 2019
Editore: Not available

Performance Analysis and Automatic Tuning of Hash Aggregation on GPUs

Autori: Viktor Rosenfeld, Sebastian Breß, Steffen Zeuch, Tilmann Rabl, Volker Markl
Pubblicato in: Proceedings of the 15th International Workshop on Data Management on New Hardware - DaMoN'19, 2019, Pagina/e 1-11, ISBN 9781-450368018
Editore: ACM Press
DOI: 10.1145/3329785.3329922

Dynamic application reconfiguration on heterogeneous hardware

Autori: Juan Fumero, Michail Papadimitriou, Foivos S. Zakkak, Maria Xekalaki, James Clarkson, Christos Kotselidis
Pubblicato in: Proceedings of the 15th ACM SIGPLAN/SIGOPS International Conference on Virtual Execution Environments - VEE 2019, 2019, Pagina/e 165-178, ISBN 9781-450360203
Editore: ACM Press
DOI: 10.1145/3313808.3313819

Maintaining Wavelet Synopses for Sliding-Window Aggregates

Autori: Ioannis Mytilinis, Dimitrios Tsoumakos, Nectarios Koziris
Pubblicato in: Proceedings of the 31st International Conference on Scientific and Statistical Database Management - SSDBM '19, 2019, Pagina/e 73-84, ISBN 9781-450362160
Editore: ACM Press
DOI: 10.1145/3335783.3335793

Exploiting high-performance heterogeneous hardware for Java programs using graal

Autori: James Clarkson, Juan Fumero, Michail Papadimitriou, Foivos S. Zakkak, Maria Xekalaki, Christos Kotselidis, Mikel Luján
Pubblicato in: Proceedings of the 15th International Conference on Managed Languages & Runtimes - ManLang '18, 2018, Pagina/e 1-13, ISBN 9781-450364249
Editore: ACM Press
DOI: 10.1145/3237009.3237016

Docker-Sec: A Fully Automated Container Security Enhancement Mechanism

Autori: Fotis Loukidis-Andreou, Ioannis Giannakopoulos, Katerina Doka, Nectarios Koziris
Pubblicato in: 2018 IEEE 38th International Conference on Distributed Computing Systems (ICDCS), 2018, Pagina/e 1561-1564, ISBN 978-1-5386-6871-9
Editore: IEEE
DOI: 10.1109/icdcs.2018.00169

Efficient window aggregation with general stream slicing

Autori: Traub, Jonas; Grulich, Philipp; Cuéllar, Alejandro Rodríguez; Breß, Sebastian; Katsifodimos, Asterios; Rabl, Tilmann; Markl, Volker
Pubblicato in: 22nd International Conference on Extending Database Technology, Numero 2019 March, 2019
Editore: Open Proceedings
DOI: 10.5441/002/edbt.2019.10

Efficient SIMD Vectorization for Hashing in OpenCL

Autori: Behrens, Tobias, Viktor Rosenfeld, Jonas Traub, Sebastian Breß, and Volker Markl
Pubblicato in: Proceedings of the 21st International Conference on Extending Database Technology (EDBT), 2018, ISBN 978-3-89318-078-3
Editore: OpenProceedings

Efficient k-means on GPUs

Autori: Clemens Lutz, Sebastian Breß, Tilmann Rabl, Steffen Zeuch, Volker Markl
Pubblicato in: Proceedings of the 14th International Workshop on Data Management on New Hardware - DAMON '18, 2018, Pagina/e 1-3, ISBN 9781-450358538
Editore: ACM Press
DOI: 10.1145/3211922.3211925

DERP: A Deep Reinforcement Learning Cloud System for Elastic Resource Provisioning

Autori: Constantinos Bitsakos, Ioannis Konstantinou, Nectarios Koziris
Pubblicato in: 2018 IEEE International Conference on Cloud Computing Technology and Science (CloudCom), 2018, Pagina/e 21-29, ISBN 978-1-5386-7899-2
Editore: IEEE
DOI: 10.1109/cloudcom2018.2018.00020

Towards practical heterogeneous virtual machines

Autori: James Clarkson, Juan Fumero, Michail Papadimitriou, Maria Xekalaki, Christos Kotselidis
Pubblicato in: Conference Companion of the 2nd International Conference on Art, Science, and Engineering of Programming - Programming'18 Companion, 2018, Pagina/e 46-48, ISBN 9781-450355131
Editore: ACM Press
DOI: 10.1145/3191697.3191730

Challenges and Proposals for Enabling Dynamic Heterogeneous Execution of Big Data Frameworks

Autori: Maria Xekalaki, Juan Fumero, Christos Kotselidis
Pubblicato in: 2018 IEEE International Conference on Cloud Computing Technology and Science (CloudCom), 2018, Pagina/e 335-341, ISBN 978-1-5386-7899-2
Editore: IEEE
DOI: 10.1109/cloudcom2018.2018.00070

Using compiler snippets to exploit parallelism on heterogeneous hardware: a Java reduction case study

Autori: Juan Fumero, Christos Kotselidis
Pubblicato in: Proceedings of the 10th ACM SIGPLAN International Workshop on Virtual Machines and Intermediate Languages - VMIL 2018, 2018, Pagina/e 16-25, ISBN 9781-450360715
Editore: ACM Press
DOI: 10.1145/3281287.3281292

The Vision of a HeterogeneRous Scheduler

Autori: Ioannis Mytilinis, Constantinos Bitsakos, Katerina Doka, Ioannis Konstantinou, Nectarios Koziris
Pubblicato in: 2018 IEEE International Conference on Cloud Computing Technology and Science (CloudCom), 2018, Pagina/e 302-307, ISBN 978-1-5386-7899-2
Editore: IEEE
DOI: 10.1109/cloudcom2018.2018.00065

Analyzing efficient stream processing on modern hardware

Autori: Steffen Zeuch, Bonaventura Del Monte, Jeyhun Karimov, Clemens Lutz, Manuel Renz, Jonas Traub, Sebastian Breß, Tilmann Rabl, Volker Markl
Pubblicato in: Proceedings of the VLDB Endowment, Numero 12/5, 2019, Pagina/e 516-530, ISSN 2150-8097
Editore: VLDB Endowment
DOI: 10.14778/3303753.3303758

Efficient and Scalable k‑Means on GPUs

Autori: Clemens Lutz, Sebastian Breß, Tilmann Rabl, Steffen Zeuch, Volker Markl
Pubblicato in: Datenbank-Spektrum, Numero 18/3, 2018, Pagina/e 157-169, ISSN 1618-2162
Editore: Springer Berlin Heidelberg
DOI: 10.1007/s13222-018-0293-x

Generating custom code for efficient query execution on heterogeneous processors

Autori: Sebastian Breß, Bastian Köcher, Henning Funke, Steffen Zeuch, Tilmann Rabl, Volker Markl
Pubblicato in: The VLDB Journal, Numero 27/6, 2018, Pagina/e 797-822, ISSN 1066-8888
Editore: Springer Verlag
DOI: 10.1007/s00778-018-0512-y

Risultati finali

Pubblicazioni

Condividi questa pagina Condividi questa pagina sui social network

Scarica Scarica il contenuto della pagina