Add a section explaining server configurations

docs/source/user_guide/dpf_concepts.rst

@@ -9,6 +9,7 @@ DPF concepts
    concepts.rst
    waysofusing.rst
    stepbystep.rst
+   server_types.rst
 
 
 .. card-carousel:: 2
@@ -36,3 +37,11 @@ DPF concepts
        :text-align: center
 
        .. image:: ../images/drawings/checklist.png
+
+    .. card:: Using DPF: locally/remotely
+       :link: user_guide_server_types
+       :link-type: ref
+       :width: 25%
+       :text-align: center
+
+       .. image:: ../images/drawings/using-dpf.png

docs/source/user_guide/index.rst

@@ -15,7 +15,7 @@ This section has the following goals:
  - Provide simple how-tos for tackling most common use cases.
 
 Other sections of this guide include :ref:`ref_api_section`, :ref:`ref_dpf_operators_reference`,
-and :ref:`sphx_glr_examples`.
+and :ref:`_gallery`.
 
 .. include::
    dpf_concepts.rst

docs/source/user_guide/server_types.rst

@@ -0,0 +1,38 @@
+.. _user_guide_server_types:
+
+===========================
+Client-server communication
+===========================
+
+DPF is based on a client-server architecture.
+
+The DPF Server acts as a server, while the different client APIs (CPython, IronPython, C++...)
+act as clients connecting to it.
+PyDPF is for example a term encompassing both the CPython client
+(the Python packages such as ansys-dpf-core or ansys-dpf-post, available on PyPI)
+and the IronPython client (available within Ansys Mechanical).
+
+The communication logic with a DPF server is defined when starting it using
+an instance of the :class:`ServerConfig <ansys.dpf.core.server_factory>` class.
+Different predefined server configurations are available in DPF,
+each answering a different use-case
+(See the :class:`AvailableServerConfigs <ansys.dpf.core.server_factory>` class).
+
+- The :class:`GrpcServer <ansys.dpf.core.server_types>` configuration is available starting with server version 4.0 (Ansys 2022R2).
+  It allows for remote connections to a DPF Server across a network by telling the client
+  to communicate with this server via the gRPC communication protocol.
+  Although it can be used to communicate with a DPF server running on the same local machine,
+  in that case the next configuration should be preferred.
+- The :class:`InProcess <ansys.dpf.core.server_types>` configuration is available starting with server version 4.0 (Ansys 2022R2).
+  It indicates to the client that a DPF server is installed on the local machine, enabling direct calls
+  to the server binaries from within the client's own Python process.
+  This removes the need to copy and send data between the client and server, and makes calls
+  to the server functionalities much faster as well as using less memory.
+- The :class:`LegacyGrpcServer <ansys.dpf.core.server_types>` configuration is the only one available for server versions below 4.0
+  (Ansys 2022R1, Ansys 2021R2 and Ansys 2021R1).
+  The client communicates with a local or remote DPF server via the gRPC communication protocol.
+
+For DPF with Ansys 2023 R1 and newer, the default configuration is set to :class:`InProcess <ansys.dpf.core.server_types>`,
+meaning that servers are launched on the local machine.
+To launch a DPF server on a remote machine and communicate with it using gRPC, one should use
+the :class:`GrpcServer <ansys.dpf.core.server_types>` configuration as shown in :ref:`_ref_server_types_example`.