.. vim: set fileencoding=utf-8 :

.. Copyright (c) 2016 Idiap Research Institute, http://www.idiap.ch/          ..
.. Contact: beat.support@idiap.ch                                             ..
..                                                                            ..
.. This file is part of the beat.web module of the BEAT platform.             ..
..                                                                            ..
.. Commercial License Usage                                                   ..
.. Licensees holding valid commercial BEAT licenses may use this file in      ..
.. accordance with the terms contained in a written agreement between you     ..
.. and Idiap. For further information contact tto@idiap.ch                    ..
..                                                                            ..
.. Alternatively, this file may be used under the terms of the GNU Affero     ..
.. Public License version 3 as published by the Free Software and appearing   ..
.. in the file LICENSE.AGPL included in the packaging of this file.           ..
.. The BEAT platform is distributed in the hope that it will be useful, but   ..
.. WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY ..
.. or FITNESS FOR A PARTICULAR PURPOSE.                                       ..
..                                                                            ..
.. You should have received a copy of the GNU Affero Public License along     ..
.. with the BEAT platform. If not, see http://www.gnu.org/licenses/.          ..


.. _administratorguide-hardware_guidelines:

Deployment Guidelines
=====================

In this section, we evaluate three common scenarios in which the BEAT platform
could be deployed. These scenarios were picked so to maximize the installation
range and serve as a reference to potential BEAT deployers.


.. _administratorguide-hardware_guidelines-single-machine:

Single machine
--------------

In this scenario, every module of the Figure
":ref:`administratorguide-architecture-hardware`" is supposed to run on one
single, possibly multi-core, machine. This is the configuration that might be
used to develop the platform and demonstrate it when no access to the
Internet is available.

The performance of this installation will depend on the number of CPU cores
that can be dedicated to the processing nodes as no network connectivity will
be required.


.. _administratorguide-hardware_guidelines-distributed-nodes:

Distributed processing nodes
----------------------------

.. _administratorguide-usecaseanalysis-distributed-nodes:
.. figure:: img/hardware-distributed-nodes.*
   :width: 80%

   Distributed processing nodes


In this scenario, depicted in
:numref:`administratorguide-usecaseanalysis-distributed-nodes`, every
processing node is running on a dedicated machine.

One computer acts as the web front-end, handling all the incoming requests from
the users, scheduling the jobs on the processing nodes and sending the results
to the users.

All the machines have access to the Cache, hosted on a separate file system.
The files needed to serve the website and to store the database are located
on the same machine than the Web Server and the Database Server.


.. _administratorguide-hardware_guidelines-distributed-architecture:

Load-balanced distributed architecture
--------------------------------------

.. _administratorguide-usecaseanalysis-load-balancing:
.. figure:: img/hardware-load-balancing.*

   Load-balanced distributed architecture

In this scenario, depicted in
:numref:`administratorguide-usecaseanalysis-load-balancing`, every module (web
server, database server, scheduler, processing node) is running on a dedicated
machine.

A load-balancing server (possibly a dedicated hardware one) distributes the
incoming requests over *M* web servers, which allows a great number of clients
to use the platform at the same time. The platform uses a replication mechanism
to distribute the load on the database over *N* database servers. *P*
processing nodes are available to let the Scheduler optimize the execution of
several experiments in parallel.

All the machines (but the Database servers) have access to the Data Cache,
hosted on a separate file system. Each web server has a local copy of the
website. Each Database server has a local copy of the database (kept in sync
via a replication mechanism).