176
5 Complex Reactive Applications: A Forward Look to Open Science
(personal computers (PC) and workstations) as local networked resources having
the clear advantage of narrowing the distance between compute resources and users.
First of all, this has resulted into a specialization of local resources for user-specific
needs thanks to user-oriented operating systems. Among these, Linux (born in 1991
and operating under the GPL licence based on the software of the GNU project).
Linux became very popular not only because of its ability to emulate terminals but
also because of its ability to read and write files from/to disks and to act as a true
kernel able to handle operating systems.
Further progress was made by enabling Linux to execute the X server and to provide an integrated system for graphic interfaces. At present Linux has become the
preferred operating system for servers in production environments and embedded
devices. Moreover, Linux has also a strong presence in the market of scientific desktops. As a matter of fact, Linux is the most popular operating system for executing
Apache, MySQL and PHP, the software grounding most of the web servers worldwide and has developed as well desktop environments interfaces similar to those of
Microsoft Windows and Mac OS X closer to the needs of the users.
The combination of PC-like user friendliness with high network connectivity
has enhanced the possibility of clustering remote local platforms and fostered the
development of high-throughput computing (HTC).
2
At the end of the mentioned three subsequent EGEE projects, a European HTC
powerful distributed platform (named EGI, the European Grid Infrastructure managed by EGI.eu [125]) was established in order to coordinate a large number of geographically dispersed compute resources connected over the public network through
the use of appropriate middleware and tools (see Fig. 5.10). As a result, GC has
become an important asset of the European scientific community enabling the concurrent execution (over hundreds of thousands of processors) of several distributed
programs for applications made of decoupled or loosely coupled tasks.
An innovative feature of grid computing is the possibility of setting new goals
to scientific research and technological applications by aggregating a large number
of highly dispersed and heterogeneous small size computers to the HPC ones (like
PRACE for the EU and XSEDE for the US) made of several millions of cores and
large storages. The main problem in this is the distance between the policies adopted
by the management of large-scale compute facilities and the expectation of a large
fraction of the Molecular science users (especially those having a high activity of
design and development of innovative codes). The compute time allotment policy
2 High-throughput computing (HTC) refers to machines exhibiting an efficient execution of a large
number of loosely-coupled tasks. HTC systems are independent sequential jobs that can be individually scheduled on many different computing resources across multiple administrative boundaries.
HTC systems achieve this using various grid computing (GC) technologies and techniques. In
Europe, a strong impulse to HTC has been given by the last Framework Programmes, especially as
a support to the High Energy Physics transnational community initiatives, by funding several international collaborative projects like DATATAG (http://datatag.web.cern.ch/datatag/), EGEE-I-II-III
(http://www.egee.eu), WLCG (http://wlcg.web.cern.ch/), EGI-Inspire (https://www.egi.eu/about/
egi-inspire/), and EGI-Engage (https://www.egi.eu/about/egi-engage/).
5 Complex Reactive Applications: A Forward Look to Open Science
(personal computers (PC) and workstations) as local networked resources having
the clear advantage of narrowing the distance between compute resources and users.
First of all, this has resulted into a specialization of local resources for user-specific
needs thanks to user-oriented operating systems. Among these, Linux (born in 1991
and operating under the GPL licence based on the software of the GNU project).
Linux became very popular not only because of its ability to emulate terminals but
also because of its ability to read and write files from/to disks and to act as a true
kernel able to handle operating systems.
Further progress was made by enabling Linux to execute the X server and to provide an integrated system for graphic interfaces. At present Linux has become the
preferred operating system for servers in production environments and embedded
devices. Moreover, Linux has also a strong presence in the market of scientific desktops. As a matter of fact, Linux is the most popular operating system for executing
Apache, MySQL and PHP, the software grounding most of the web servers worldwide and has developed as well desktop environments interfaces similar to those of
Microsoft Windows and Mac OS X closer to the needs of the users.
The combination of PC-like user friendliness with high network connectivity
has enhanced the possibility of clustering remote local platforms and fostered the
development of high-throughput computing (HTC).
2
At the end of the mentioned three subsequent EGEE projects, a European HTC
powerful distributed platform (named EGI, the European Grid Infrastructure managed by EGI.eu [125]) was established in order to coordinate a large number of geographically dispersed compute resources connected over the public network through
the use of appropriate middleware and tools (see Fig. 5.10). As a result, GC has
become an important asset of the European scientific community enabling the concurrent execution (over hundreds of thousands of processors) of several distributed
programs for applications made of decoupled or loosely coupled tasks.
An innovative feature of grid computing is the possibility of setting new goals
to scientific research and technological applications by aggregating a large number
of highly dispersed and heterogeneous small size computers to the HPC ones (like
PRACE for the EU and XSEDE for the US) made of several millions of cores and
large storages. The main problem in this is the distance between the policies adopted
by the management of large-scale compute facilities and the expectation of a large
fraction of the Molecular science users (especially those having a high activity of
design and development of innovative codes). The compute time allotment policy
2 High-throughput computing (HTC) refers to machines exhibiting an efficient execution of a large
number of loosely-coupled tasks. HTC systems are independent sequential jobs that can be individually scheduled on many different computing resources across multiple administrative boundaries.
HTC systems achieve this using various grid computing (GC) technologies and techniques. In
Europe, a strong impulse to HTC has been given by the last Framework Programmes, especially as
a support to the High Energy Physics transnational community initiatives, by funding several international collaborative projects like DATATAG (http://datatag.web.cern.ch/datatag/), EGEE-I-II-III
(http://www.egee.eu), WLCG (http://wlcg.web.cern.ch/), EGI-Inspire (https://www.egi.eu/about/
egi-inspire/), and EGI-Engage (https://www.egi.eu/about/egi-engage/).