GILDA
Grid INFN Virtual Laboratory for Dissemination Activities
Roberto Barbera
1,2
, Leandro N. Ciuffo
1
, Emidio Giorgio
1
Antonio Calanducci
1
and Valeria Ardizzone
1
1
Istituto Nazionale di Fisica Nucleare, Sez. di Catania, Via S. Sofia 64 – 95123, Catania, Italy
2
Dipartimento di Fisica e Astronomia dell’Università di Catania, Via S. Sofia 64 – 95123, Catania, Italy
Keywords: Grid computing, Training, Virtual Laboratory, gLite, e-infrastructure.
Abstract: The Grid INFN virtual Laboratory for Dissemination Activities (GILDA) is a fully fledged Grid test-bed
devoted to training and outreach activities. Open to anyone who wants to have its first hands-on experience
with grid systems, GILDA has been adopted as the official training infrastructure by many Grid projects all
around the world. All services, tools and materials produced in the past tutorials can be freely used by
anyone who wants to learn and teach Grid technology. Additionally, through a set of applications ported on
its Grid Infrastructure, developers can identify components and learn by examples how to “gridify” their
applications. This work presents the main features of such training infrastructure.
1 INTRODUCTION
Launched in 2004 by the Italian National Institute
for Nuclear Physics (INFN), GILDA (the Grid INFN
virtual Laboratory for Dissemination Activities) is a
fully fledged Grid test-bed devoted to dissemination
activities. This infrastructure is open to anyone who
wants to have its first hands-on experience with Grid
systems. Actually, GILDA can be an important tool
for at least three main categories of users:
1. Grid newcomers – people willing to start
learning how to use a grid infrastructure;
2. Grid application developers – Just like
"webifying" applications to run on a web
browser, grid users need to "gridify" their
applications to run on a Grid. Through a set of
applications ported on GILDA, developers can
identify components and learn by examples how
to “gridify” their own applications;
3. Tutors – GILDA has been developed keeping
training and education in mind. Thus, all
services, tools and materials produced in the past
tutorials can be freely used by anyone who wants
to learn or teach Grid technology.
Indeed, GILDA has been adopted as the official
training platform by several former and current Grid
projects, such as EGEE/EGEE-II/EGEE-III
(www.eu-egee.org), EELA/EELA-2 (www.eu-
eela.eu), EU-IndiaGRID (www.euindiagrid.org),
EUMEDGRID (www.eumedgrid.org),
EUChinaGRID (www.euchinagrid.org), ICEAGE
(www.iceage-eu.org), OMII-EU (http://omii-
europe.org), BEinGRID (www.beingrid.eu) and
many others, becoming a "de facto" standard
training-Infrastructure (referred hereafter as t-
infrastructure) in Europe and in several other parts
of the world for dissemination of Grid Computing
technology.
GILDA objectives can be summarized as
follows: (i) to raise awareness of Grid Computing
benefits; (ii) to provide customized formats for
dissemination events, according to the skills of
attendants; (iii) to facilitate appropriate free on-line
content and services for training purposes; and (iv)
to encourage the use of a complete t-infrastructure
by new communities.
This article aims at presenting an overview of
GILDA facilities as well as to invite the reader to try
such t-infrastructure.
2 BACKGROUND
Computational and storage limitations are key issues
for organizations that depend on computation-
intensive applications. Such organizations are
frequently affected by market pressures to reduce
deployment time and maintenance costs. Hence,
273
Barbera R., N. Ciuffo L., Giorgio E., Calanducci A. and Ardizzone V. (2009).
GILDA - Grid INFN Virtual Laboratory for Dissemination Activities.
In Proceedings of the First International Conference on Computer Supported Education, pages 273-278
DOI: 10.5220/0001963402730278
Copyright
c
SciTePress
they may be looking for ways to improve the
effectiveness of their infrastructure or their business
processes through transformation, or seeking
opportunities for innovation that will benefit the
business. Grid is not the answer by itself, but in
many of these cases, it can certainly play an
important role, allowing immediate productivity and
benefits and giving more choices and control on how
to purchase and leverage IT power for competitive
advantage. The Grid vision is to expand parallel and
distributed computation, providing a virtualization
of heterogeneous compute and data resources,
supporting security policy based resource allocation
and prioritization. The Grid is ideal for any
applications requiring excellent performance and
scalability for their compute-intensive processes
(e.g., Monte Carlo simulation, engineering CAD
simulations, protein modeling, 3D rendering,
computational archaeology investigations, etc.).
In this context, the GILDA Project was born with
the aim of offering a one-of-a-kind service for those
interested in testing the Grid, using gLite (2008) and
the EGEE infrastructure with their own systems.
GILDA offers either basic experiences through the
“Try the Grid” (EGEE, 2006) walkthrough in
minutes or intensive and in-depth training by helping
users willing to develop applications to be ported
into a Grid environment.
3 DISSEMINATION TOOLS
The main objectives of GILDA activity around the
world are to encourage and help new and existing
communities to support them for improvement or
migration of their applications to Grid
infrastructures, to accelerate the adoption of Grid
technologies, and to increase the satisfaction of
those currently using the Grid services through the
communities’ feedback. Training activities are a key
component of the knowledge dissemination process,
ensuring that all users fully understand the
characteristics of the offered Grid services and that
they have enough expertise to properly use the
available Grid infrastructure. In order to achieve the
main objectives, several dissemination instruments
are used. A brief description of these instruments is
presented below.
3.1 Tutorials for Applications
Developers
Porting an application into a Grid environment has
never been an easy task. In order to enable
application developers to get used to the main Grid
functionalities, an advanced tutorial has been created
fully dedicated to teach them how to “gridify” their
applications. Such training events are the perfect
scenario to put application experts in tight
collaboration with Grid experts.
Hence, the GILDA Team has acquired a good
experience in transferring knowledge and know-how
to new communities by helping them to integrate
their applications into several Grid projects’
infrastructures.
3.2 Tutorials for Sysadmin
Tutorials for grid system administrators are
organized by the GILDA Team to meet the needs of
computer centers interested in joining the GILDA
test-bed or other Grid infrastructures. It is also worth
mentioning that it is possible to have your own
GILDA-like stand-alone Grid installed in your
institution. In fact, it is a common practice adopted
to support the organization of long training events,
such as Grid Schools or graduation courses.
Usually, for a novice user, installing a new Grid
service and configuring it properly are not
straightforward tasks. In that scenario, a step-by-step
guide is desired to help user during the
troubleshooting stage. The appropriate
documentation on site installation can be found at
(GILDA, 2004d).
The sysadmin tutorials are divided into three
parts: a theoretical part, a practical one and the
hands-on section. In each presentation, the GILDA
tutors use slides to show the main steps necessary to
install and configure a Grid element. Then, in the
hands-on section all participants try to install a Grid
element by themselves, assisted by tutors.
Training material is composed of multimedia
slides, a Wiki website and a set of "Virtual
Services", detailed below.
The most important stage during the preparation
of a tutorial is to set up the machines used for the
grid node installation: to make it easy, a
virtualization technique such as VMware® is
adopted in all organized tutorials. In this way, all
participants have their own Virtual Machines on
which they can work with minimum efforts required
from the system administrators.
3.3 GILDA Wiki
The GILDA Wiki (2008) has been created mainly
with the purpose of documenting and organizing the
huge amount of training material produced so far. Its
CSEDU 2009 - International Conference on Computer Supported Education
274
contents are freely available on the Web for every
interested user.
Moreover, this site is not a simple on-line
documentation repository, but an important
collaborative tool used by the whole project team.
Thus, all registered user can contribute feeding the
site with useful additional training material.
The Wiki site is an information source for three
target audiences: users, site administrators and
application developers. Regarding the users’ content,
the available material is subdivided into three
different complexity levels (Basic, Medium and
Advanced), while the site administrators area
consists in step-by-step procedures to install and
configure Grid services. Finally, the developers
section presents the middleware functionalities that
can be integrated into the source code using the API.
3.4 Support System
Through a ticket-based support system (GILDA,
2004a), users can send their questions to the GILDA
Team and get customized answers to their issues.
The GILDA support system is an important
communication tool aimed at helping anyone facing
problems on using or installing a GILDA site.
3.5 Virtual Services
Through the use of the virtualization technique, it is
possible to carry out all Grid elements. This
instrument has many benefits, such as increasing
services’ portability and reducing both the time to
put a site in operation and the number of real
machines required. The complete list of available
virtual services is available at (GILDA, 2004b).
4 THE GILDA STACK
Four main layers can be identified in the GILDA
environment, as depicted in Figure 1. At the lowest
level, dedicated hardware resources, such as
computers and digital networks running the required
operating systems and protocols, as well as local
resources, which can be either physical or virtual,
such as emulations of computers and operating
systems on which a Grid is to be built. At the second
level resources are made accessible by software
services (middleware, databases, APIs, course-
specific applications). At the third level, guides and
cookbooks suggesting how to build the t-
Infrastructure such that everyone can replicate it if
he/she wishes or needs it.
Finally, on top there should be matching
compositions of: rubrics and instruction texts,
example data sets modified in scale and content to
suit educational goals and data use policies,
exemplar applications, problem sets, software to
support exercises, and everything which makes the
infrastructure profitable by students.
Figure 1: The GILDA stack.
5 THE GILDA TEST-BED
The GILDA test-bed is maintained on a “best-effort”
basis. The computational resources usually comes
from institutions located in Europe, Asia or Latin
America. Therefore new grid sites, using
heterogeneous hardware, are frequently joining the t-
Infrastructure, just like a “real world” Grid
environment.
The test-bed itself is made up of all the
components of a larger real Grid infrastructure,
including services, resources and monitoring tools.
To allow the use of the test-bed, it also features a
Virtual Organization (VO) and a real Certification
Authority (CA) that grants two-week certificates for
the test use of the GILDA infrastructure. In addition,
it runs the latest production (and stable) version of
the gLite middleware (2008) developed in the
context of the EGEE project.
In order to ease the access to its infrastructure,
GILDA also provides a grid portal called GENIUS
(GILDA,2004c), where different usages are
supplied, such as basic content for novice users and
full featured ones for more in depth tutorials and
demonstrations. At a more advanced level, GILDA
offers a one-of-a-kind service for those interested in
testing the Grid and gLite with their own software,
offering a more intensive and in-depth introduction
to the Grid. In a couple of weeks, an user can be
trained in Grid usage, his/her software can be
modified in order to run on the Grid environment
and finally, a GENIUS service can be created to
increase the visibility of the work that has been
done.
GILDA - Grid INFN Virtual Laboratory for Dissemination Activities
275
6 VIRTUOUS CYCLE
By its nature, GILDA is one of the key enablers of
the “virtuous cycle” to attract and support new
communities. Its workflow, depicted in Figure 2, is
described a below:
1. Driven either by the dissemination events or
scientific papers where GILDA is cited, a novice
user can get the feeling of what Grid Computing
is and which kind of applications can run on a
Grid infrastructure by exploring the GILDA web
portal or using the Grid Demonstrator interface
(GILDA,2006).
2. Following up the dissemination activities,
institutions usually request the organization of
Grid training events based on the GILDA t-
infrastructure. An interested user, participating
in such tutorial or induction course, can go
through all the mandatory procedure of the
request of a personal digital certificate and
subscription to a Virtual Organization and then
use the Grid Tutor machine (GILDA,2004c);
3. After participating in a training event or even
doing self-training by following the GILDA wiki
pages, many users try to port their own
applications on the test-bed. In a smaller scale,
these users will face all problems of interfacing
the Grid services available before entering in a
huge production quality e-Science infrastructure;
4. Various applications from different communities
ported on GILDA can be either deployed on
large e-Infrastructures or incorporated into the
Grid Demonstrator, enriching the portfolio of
examples that can be demonstrated to new
people.
Figure 2: “Virtuous cycle” of applications on GILDA.
7 APPLICATIONS
The Grid environment requires many new skills both
for scientists that need to learn how to work on it
and for application developers that have to learn
how to write and optimize codes to properly interact
with the available Grid services and computational
resources. As a consequence, disciplinary support is
also essential to carry out Grid knowledge
management and Grid education, both generic and
towards specific application domains. In GILDA a
set of "Case Studies" of integrated applications has
been created, so that scientist or software developers
can explore and compare different approaches for
the integration process of their application on a Grid
infrastructure.
Currently, GILDA applications portfolio covers a
large range of different research domains. A short
list of applications successfully ported into GILDA
test-bed is presented below.
7.1 GATE
Gate is a C++ platform based on the Monte Carlo
Geant4 toolkit (CIRRONE,2005). It has been
designed to model nuclear medicine applications,
such as PET (Positron Emission Tomography) and
SPECT (Single Photon Emission Computed
Tomography) in the context of the OpenGATE
collaboration (GATE). Its functionalities combined
to its ease-of-use make this platform also impressive
for radiotherapy and brachytherapy treatment
planning.
7.2 Hadron Therapy
HadronTherapy simulates the beam line and
particles detectors used in the proton-therapy facility
CATANA (Centro di AdroTerapia e Applicazioni
Nucleari Avanzate), operational at INFN-LNS.
Here, a 62 MeV (62 million of electron volts) proton
beam, accelerated by a superconducting cyclotron, is
used for the treatment of some kind of eye cancer.
7.3 gMOD
gMOD is a new application offering a content-on-
demand service. A user can browse a catalog of
multimedia contents and request one to be streamed
in real time to his/her workstation using the power of
the Grid. Movies can also be chosen by querying an
underlying metadata catalog using one or more
attributes. The role feature of the Virtual
Organization Membership Services (VOMS) is
exploited to define normal users and catalog
managers. gMOD represents a very interesting use
case for the use of Grid both in entertainment and
distributed digital content management systems such
as the e-learning ones.
CSEDU 2009 - International Conference on Computer Supported Education
276
7.4 Volcano Sonification
Current knowledge of volcanic eruptions does not
yet allow scientists to predict future eruptions. This
application represents an attempt to put the scientific
community one step closer to the prediction asset by
means of the sonification of volcano seismograms.
Thus, the translation of the patterns of Mount Etna
(Italy) and Mount Tungurahua (Ecuador) volcanic
behaviors into sound waves has been carried.
Data sonification is currently used in several fields
and for different purposes: science and engineering,
education and training. It acts mainly as data
analysis and interpretation tool. Figure 3 illustrates
its working scheme.
Figure 3: Volcano Sonification – input / output scheme.
8 FORTHCOMING FEATURES
The t-infrastructure provided by GILDA will be
continuously used by many EU funded projects,
since several project proposals submitted to the 7th
Framework Programme calls expressed their
intention to use GILDA facilities as part of their
training program.
The GILDA hardware expansion plan foreseen
the integration of new 64-bits worker nodes by the
end of 2008 (up to now, only 32-bits nodes are
available). This will lead the availability of new
libraries often requested by application developers
and the general industry community such as MPI-2.
Additionally, the growing number of users has
encouraged the GILDA team to recently announce
the incorporation of some new features. These
features are:
8.1 gLibrary
gLibrary (CALANDUCCI,2007) an extensible,
robust, secure and easy-to-use system to handle
digital libraries in a Grid infrastructure. It offers an
intuitive Web interface to browse the available
entries, and thanks to its powerful “iTunes-like”
searching capabilities based on attribute filters,
finding a library asset is just a matter of seconds.
gLibrary is a flexible system that can be used for
different purposes and for different communities that
can easily adopt this framework to build their own
digital libraries defining types and categories
according to their needs.
8.2 Secure Storage Service
This service was carried out by UNICO S.r.l.
(http://www.unicosrl.it) in close collaboration with
INFN in the context of the TriGrid VL Project
(http://www.trigrid.it). This service provides the
users with a set of tools for storing confidential data
(e.g., medical or financial data) in a secure way and
in an encrypted format on the Grid storage elements.
The data stored in this way are accessible and
readable by authorized users only. Moreover, it
solves the insider abuse problem preventing also the
administrators of the storage elements from
accessing the confidential data in a clear format.
8.3 GRelC
The Grid Database Access Service (GRelC) (2007)
aims to provide a set of advanced data grid service
to transparently, efficiently and securely manage
databases in a Grid environment (FIORE,2007).
This framework, developed by University of Lecce –
Italy and SPACI Consortium (The Italian Southern
Partnership for Advanced Computational
Infrastructures), is currently used to support
bioinformatics experiments on distributed and huge
data banks. The framework provides a uniform
access interface to access and interact with
heterogeneous DBMS such us PostgreSQL, MySQL,
Oracle, DB2, SQLite, etc.
9 CONCLUSIONS
GILDA represents a very valuable tool in the Italian
INFN Grid Project (http://grid.infn.it) as well as in
several Grid projects around the world. Indeed, its t-
infrastructure has been used so far in more than 300
induction courses in 53 different countries all over
GILDA - Grid INFN Virtual Laboratory for Dissemination Activities
277
the world. Altogether, more than 13000 certificates
were issued by the GILDA Certification Authority,
and at the time this article was written, more than a
thousand users were currently registered with the
GILDA Virtual Organization.
Through GILDA a broad range of users are able
to get quick and easy access to a “real world” Grid
environment. Furthermore, GILDA facilities are
open to any organizations or educational institutions
who want to adopt it in their Grid training programs.
In that sense, the main business benefits of adopting
GILDA can be summarized as follows:
1. It is free;
2. No costs to leverage Grid resources are required
to use it;
3. It is the first place where a user can takes
experience on Grid Computing;
4. It provides an affordable and pleasant experience
in Grid training;
5. It provides a good infrastructure where business
applications can try “gridifying” operation
exploiting the GILDA Team support.
6. It improves local research activity quality
providing more power computation and storage
resources;
7. It contributes to promote international
cooperation and partnership.
ACKNOWLEDGEMENTS
The GILDA Team would like to thank all the
partners (project and institutions) that collaborate on
proving resources, contents, services and human
effort to this project.
REFERENCES
Barbera, R., Falzone A. & Rodolico A. (2003, March).
The GENIUS Grid Portal. Paper presented at
Conference for Computing in High Energy and
Nuclear Physics (CHEP 2003), La Jolla, California,
USA.
Calanducci,A., Cherubino,C., Ciuffo,L.N. & Scardaci,D.
(2007). “A Digital Library Management System for the
Grid”, In: Fourth International Workshop on
Emerging Technologies for Next generation GRID
(ETNGRID 2007) at 16th IEEE International
Workshops on Enabling Technologies: Infrastructures
for Collaborative Enterprises (WETICE-2007),
GET/INT Paris, France.
Cirrone, G.A.P., Cuttone, G. et al. (2005) Implementation
of a new Monte Carlo-GEANT4 Simulation tool for
the development of a proton therapy beam line and
verification of the related dose distributions. In: IEEE
Transactions on Nuclear Science, Vol 52, Issue 1, Part
1, pp.262-265.
EGEE. (2006). Try the Grid. Retrieved February 8, 2009,
from http://egee2.eu-egee.org/try-the-grid
Fiore S., Cafaro M. & Aloisio G. (2007, June) GRelC
DAS: a Grid-DB Access Service for gLite Based
Production Grids. Paper presented at the Fourth
International Workshop on Emerging Technologies for
Next-generation GRID (ETNGRID 2007), Paris,
France.
GATE. (n.d.). OpenGATE collaboration. Retrieved
February 8, 2009, from
http://opengatecollaboration.healthgrid.org
GILDA. (2004a). Support. Retrieved February 8, 2009,
from http://gilda-support.ct.infn.it/
GILDA. (2004b).Virtual Services. Retrieved February 8,
2009, from https://gilda.ct.infn.it/VirtualServices.html
GILDA (2004c) Grid Tutors. Retrieved February 8, 2009,
from https://gilda.ct.infn.it/grid-tutor.html
GILDA (2004d). Instructions for sites. Retrieved February
8, 2009, from https://gilda.ct.infn.it/sites.html
GILDA. (2006). Grid Demonstrator. Retrieved February
8, 2009, from https://glite-demo.ct.infn.it/
GILDA. (2008). TWiki. Retrieved February 8, 2009, from
http://grid.ct.infn.it/twiki/bin/view/GILDA/
gLite 3.0. (2008). Lightweight Middleware for Grid
Computing. Retrieved February 8, 2009, from
http://glite.web.cern.ch/glite/
GRelC. (2007). Grid Relational Catalog Project.
Retrieved February 8, 2009, from http://grelc.unile.it/
INFN Grid. (n.d.). INFN Grid - The Italian Grid
Infrastructure. Retrieved February 8, 2009, from
http://grid.infn.it/
LCG. (n.d.) LCG Middleware. Retrieved February 8,
2009, from
http://lcg.web.cern.ch/LCG/activities/middleware.html
CSEDU 2009 - International Conference on Computer Supported Education
278