EDGeS@Home

EDGeS@Home (Enabling Desktop Grids for e-Science) was a BOINC-based volunteer computing project that aimed to bridge desktop grid (DG) and service grid (SG) infrastructures, enabling the execution of scientific applications developed by the EGEE and EDGeS community on volunteer-contributed computing resources. The project was coordinated by the MTA SZTAKI Laboratory of Parallel and Distributed Systems in Hungary and was part of the broader EDGeS European Union Seventh Framework Programme (FP7) project.

EDGeS@Home
File:EDGeS logo.png EDGeS@Home project logo
File:EDGeS@Home screenshot.png EDGeS@Home project homepage (archived)
Project
Status	Completed
Category	Plasma physics, Grid computing
Compute	CPU
Requires	BOINC client
Development
Developer	MTA SZTAKI Laboratory of Parallel and Distributed Systems
Author	Péter Kacsuk, Zoltán Farkas
Sponsor	European Union FP7 (grant 211727); DEGISCO (grant 261556)
Maintainer	EDGeS team
Initial release	October 28, 2009  (17 years ago)
Completed	May 2011 (ISDEP phase)
Discontinued	ca. 2015
Software
Written in	C
Operating system	Windows, Linux, macOS
Size	~70 MB (ISDEP application)
BOINC statistics
Stats as of	November 28, 2015  (11 years ago)
Performance	1.35 GFLOPS
Active users	0
Total users	11,486
Active hosts	0
Total hosts	32,308
Analytics
RAC	0
Credit/day	0
CPU performance	1.35 GFLOPS
Metadata
Website	http://home.edges-grid.eu/home/

The project's primary production-level application was ISDEP (Integrator of Stochastic Differential Equations for Plasmas), a Monte Carlo code that simulates ion transport in magnetized plasmas for fusion energy research. EDGeS@Home also hosted several other applications at the beta (experimental) level.

History

The EDGeS (Enabling Desktop Grids for e-Science) project was an European Union FP7 infrastructure project (grant agreement no. 211727) that ran from 2008 to 2010, with the goal of building technological bridges to facilitate interoperability between desktop grids and service grids. The EDGeS@Home BOINC project was launched on 28 October 2009 as the volunteer computing component of this infrastructure.

The TJ-II stellarator at CIEMAT, Madrid, the primary fusion device studied using ISDEP on EDGeS@Home.

The EDGeS@Home desktop grid and its applications were partly supported by the follow-on DEGISCO project (Desktop Grid Initiatives for Scientific Collaboration), funded under EU FP7 grant agreement no. 261556. The experts of the International Desktop Grid Federation (IDGF) provided further support for the EDGeS@Home infrastructure, its applications, and its integration into the DEGISCO infrastructure.

A server upgrade was performed in April 2011, after which the project continued operating. The ISDEP computation phase was completed on 21 May 2011. After the ISDEP phase ended, the project briefly hosted an AutoDock application for molecular docking computations at the beta level. The project became inactive by approximately 2015, with the last statistics export occurring on 28 November 2015.

Scientific objectives

The aim of the EDGeS@Home project was to support the execution of selected and validated scientific applications developed by the EGEE and EDGeS community on volunteer computing resources. The project served as a practical demonstration of the EDGeS bridging technology, which connected BOINC-based and XtremWeb-based desktop grids with the EGEE service grid infrastructure, creating a seamless pool of over 250,000 processors from desktop grids combined with approximately 150,000 processors from service grids.

ISDEP

EDGeS 3G Bridge architecture diagram

The EDGeS 3G Bridge architecture connecting BOINC desktop grids with the EGEE service grid, enabling bidirectional job execution.

The principal application hosted on EDGeS@Home was ISDEP (Integrator of Stochastic Differential Equations for Plasmas), a Monte Carlo code developed under the collaboration between the National Fusion Laboratory at CIEMAT (Madrid), the Institute of Biocomputation and Physics of Complex Systems (BIFI) at the University of Zaragoza, and the Complutense University of Madrid.

ISDEP solves the dynamics of a minority population of ions in a complex 3D fusion device, computing the distribution function by numerically integrating the Langevin equations equivalent to the Fokker–Planck equation. The code takes into account:

• The full three-dimensional structure of fusion devices, avoiding common approximations such as Boozer coordinates and the neoclassical ordering
• The confining electromagnetic field
• Collisions with other plasma species (ions and electrons) using the Boozer–Kuo-Petravic collision operator
• The electrostatic potential

The Monte Carlo method used by ISDEP is based on the equivalence between the Fokker–Planck and Langevin equations, which allows the code to run on distributed computing platforms without communication between nodes, achieving near-linear scaling. Each computing node integrates one or more ion trajectories independently, with statistical independence guaranteed by reading the random seed locally.

The primary fusion device studied with ISDEP on EDGeS@Home was the TJ-II stellarator at CIEMAT. ISDEP was also applied to study ion transport in ITER and the Large Helical Device (LHD) stellarator.

ISDEP was programmed in C to maximize performance and portability, and was designed from the outset to run on grid computing and volunteer computing platforms. The application download size was approximately 70 MB, with input data per workunit of 212 bytes and output data of 500–700 KB.

Other applications

In addition to ISDEP at production level, EDGeS@Home hosted several other applications at the beta (experimental) level:

• AutoDock – a molecular docking application used in biomedical research, which became the sole active application after the ISDEP phase ended in May 2011
• Applications from the broader EDGeS infrastructure, including those bridged from EGEE service grids via the 3G Bridge technology

The ISDEP application was also available through the Spanish volunteer computing project Ibercivis and its predecessor ZIVIS (Zaragoza Institutional Volunteer Computing for Investigative Science), a city-wide computing platform deployed in Zaragoza, Spain.

Bridging technology

The core technical contribution of the EDGeS project was the development of 3G Bridge (Generic Grid-to-Grid Bridge), a middleware-independent framework that enabled interoperability between desktop grids and service grids. The bridge operated in two directions:

• BOINC → EGEE bridge: A modified BOINC client that represented itself as a powerful desktop PC with many CPU cores to the EGEE Workload Management System (WMS), pulling workunits from the BOINC server and executing them on EGEE resources. This was based on the super-worker approach.
• EGEE → BOINC bridge: Allowed EGEE users to submit jobs that were transparently forwarded to and executed on connected desktop grid systems, extending the capacity of EGEE Virtual Organizations (VOs) with volunteer resources.

The production infrastructure included the SZTAKI Desktop Grid, the Extremadura Desktop Grid, and the AlmereGrid, connected to EGEE service grids through the 3G Bridge.

Volunteer participation

Volunteers participated in EDGeS@Home by installing the BOINC client and attaching to the project URL http://home.edges-grid.eu/home/. At its peak, the project reached approximately:

Metric	Value
Total users	11,486
Total hosts	32,308
Total credit	460,481,463
Peak estimated performance (48 h)	~1.57 TFLOPS
Number of countries	84
Number of teams	684

As of March 2012, over 7,000 users from 84 countries had contributed more than 17,000 computers, providing an integrated performance of approximately 2.6 TFLOPS. The project had 11,032 users and 16,385 hosts as of March 2012, with an estimated 48-hour performance of 1,352.6 GFLOPS.

Relationship to other projects

EDGeS@Home was closely related to several other grid and volunteer computing initiatives:

• Ibercivis – A Spanish volunteer computing project that also hosted the ISDEP application. EDGeS@Home served as part of a unified European grid infrastructure that included Ibercivis as a sub-grid.
• SZTAKI Desktop Grid – A BOINC-based desktop grid developed at MTA SZTAKI using the DC-API (Distributed Computing Application Programming Interface), which was a foundational component of the EDGeS infrastructure.
• ZIVIS – A city-wide volunteer computing platform in Zaragoza, Spain, that served as a precursor to Ibercivis and hosted ISDEP as its pilot application.
• AlmereGrid – A Dutch desktop grid that was part of the EDGeS production infrastructure.
• DEGISCO – The follow-on project (FP7 grant 261556) that supported the EDGeS@Home infrastructure and promoted interoperation between desktop grid and service grid infrastructures on a global scale.

Mathematical formulation

ISDEP solves the Fokker–Planck equation for a minority ion population in a magnetized plasma:

$\frac{\partial f}{\partial t}+\dot{\mathbf{𝐱}}\cdot \mathrm{\nabla }f+\dot{\mathbf{𝐯}}\cdot {\mathrm{\nabla }}_{v}f=C(f)$

where $f(\mathbf{𝐱},\mathbf{𝐯},t)$ is the ion distribution function, $\mathbf{𝐱}$ and $\mathbf{𝐯}$ are position and velocity in Cartesian coordinates, and $C(f)$ is the collision operator (Boozer–Kuo-Petravic).

By exploiting the equivalence between the Fokker–Planck equation and the Langevin equations, ISDEP integrates the stochastic differential equations:

$d\mathbf{𝐗}=\mathbf{𝐚}(\mathbf{𝐗},t)dt+\mathbf{𝐛}(\mathbf{𝐗},t)d\mathbf{𝐖}$

where $\mathbf{𝐚}$ is the drift term (deterministic motion in the electromagnetic field), $\mathbf{𝐛}$ is the diffusion coefficient (collisional effects), and $\mathbf{𝐖}$ is a Wiener process. The numerical integration uses a Runge–Kutta method upgraded to handle Gaussian noise from collisional effects.

Since individual particle trajectories are statistically independent, no inter-node communication is required during computation, making the problem embarrassingly parallel and well-suited for volunteer computing platforms.

Publications

BOINC-listed publications

The following 12 publications are listed on the BOINC publications page for EDGeS@Home:

1. Kacsuk, P., Z. Farkas, J. Kovács et al. Desktop grid in the era of cloud computing. (2015).
2. Fedak, Gilles. Contributions to Desktop Grid Computing. (2015). Habilitation à diriger des recherches.
3. Delamare, Simon, Gilles Fedak, Derrick Kondo and Oleg Lodygensky. SpeQuloS: a QoS service for hybrid and elastic computing infrastructures. Cluster Computing (2014). DOI: 10.1007/s10586-013-0283-6.
4. Visegrádi, Ádám, J. Kovács and P. Kacsuk. Efficient extension of gLite VOs with BOINC based desktop grids. Future Generation Computer Systems (2014). DOI: 10.1016/j.future.2013.10.012.
5. Marosi, Attila, J. Kovács and P. Kacsuk. Towards a volunteer cloud system. Future Generation Computer Systems (2013). DOI: 10.1016/j.future.2012.03.013.
6. Delamare, Simon, Gilles Fedak, Derrick Kondo, O. Lodygensky, P. Kacsuk, J. Kovács and F. Araujo. Advanced QoS Prototype for the EDGI Infrastructure. (2013).
7. Delamare, Simon, Gilles Fedak, Derrick Kondo, O. Lodygensky, P. Kacsuk, J. Kovács and F. Araujo. Intermediate QoS Prototype for the EDGI Infrastructure. (2013).
8. Terstyanszky, Gabor, Tamas Kiss, Tamas Kukla, Zsolt Lichtenberger, Stephen Winter, Pamela Greenwell, Sharron McEldowney and Hans Heindl. Application repository and science gateway for running molecular docking and dynamics simulations. Studies in Health Technology and Informatics (2012).
9. Kacsuk, P., Z. Farkas and Z. Balaton. EDGeS Bridge Technologies to Interconnect Service and Desktop Grids. Remote Instrumentation Services on the e-Infrastructure (2011).
10. Velasco, J. L., F. Castejón and A. Tarancón. Finite orbit width effect in ion collisional transport in TJ-II. Physics of Plasmas (2009). DOI: 10.1063/1.3126583.
11. Fedak, Gilles, Haiwu He, Oleg Lodygensky et al. EDGeS: A Bridge between Desktop Grids and Service Grids. 2008 Third ChinaGrid Annual Conference (2008). DOI: 10.1109/ChinaGrid.2008.44.
12. Antolí, B., F. Castejón, A. Giner et al. ZIVIS: A City Computing Platform Based on Volunteer Computing. (2007).

Additional publications

The following additional publications are related to the ISDEP application and the EDGeS project but are not listed on the BOINC publications page:

1. Urbah, E., P. Kacsuk, Z. Farkas et al. EDGeS: Bridging EGEE to BOINC and XtremWeb. Journal of Grid Computing 7, 335–354 (2009). DOI: 10.1007/s10723-009-9137-0.
2. Velasco, J.L., A. Bustos, F. Castejón, L.A. Fernández, V. Martín-Mayor and A. Tarancón. ISDEP: Integrator of stochastic differential equations for plasmas. Computer Physics Communications 183(9), 1877–1883 (2012). DOI: 10.1016/j.cpc.2012.04.004.
3. Bustos, A., F. Castejón, L.A. Fernández, J. García, V. Martín-Mayor, J.M. Reynolds, R. Seki and J.L. Velasco. Impact of 3D features on ion collisional transport in ITER. Nuclear Fusion 50(12), 125007 (2010). DOI: 10.1088/0029-5515/50/12/125007.
4. Bustos, A., J.M. Fontdecaba, F. Castejón, J.L. Velasco, M. Tereshchenko and J. Arévalo. Studies of the fast ion energy spectra in TJ-II. Physics of Plasmas 20(2), 022507 (2013). DOI: 10.1063/1.4793731.
5. Castejón, F., L.A. Fernández, J. Guasp, V. Martín-Mayor, A. Tarancón and J.L. Velasco. Ion kinetic transport in the presence of collisions and electric field in TJ-II ECRH plasmas. Plasma Physics and Controlled Fusion 49(6), 753 (2007). DOI: 10.1088/0741-3335/49/6/006.
6. Rivero, A. and D. Ferrer. ISDEP, a fusion application deployed in the EDGeS project. EGEE User Forum, EnterTheGrid workshop (2010).
7. Cárdenas-Montes, M., A. Emmen, A.C. Marosi et al. EDGeS: bridging Desktop and Service Grids. EGEE User Forum (2008).

See also

• BOINC
• Volunteer computing
• Ibercivis
• TJ-II stellarator
• ITER
• Stellarator
• Plasma (physics)