The Gateway Exchange Protocol (GEP) is a measurement-based publish/subscribe transport protocol used for efficiently exchanging time-series data. Through use of a command channel it can also automatically synchronize meta-data among applications. The protocol supports sending real-time and historical data at full or down-sampled resolutions. When sending historical data, the communications speed can be controlled dynamically to provide data as quickly as possible, e.g., a data download, or slowed for visualization streaming.

The wire protocol employed by GEP implements a publish/subscribe data exchange model using a simple command driven service with tightly compressed, fast binary serialization of time-series values. The protocol does not require a predefined or fixed configuration – that is, the time-series values arriving in one data packet can be different than those arriving in another. Each packet data consists of a collection of time-series values; each time-series value is a structure containing an ID, a time-stamp, a value and associated flags.

---

This is the C++ implementation of the Gateway Exchange Protocol.

This code includes functionality for both "subscriber" and "publisher" in GEP.

---

Compiling in Visual Studio

To properly compile in Visual Studio, you will need to download Boost: http://www.boost.org/users/download/

By default, the GSF Time Series Platform Library project configuration adds an additional include directory for the Boost libraries in a parallel location to the Grid Solutions Framework project in a folder called "boost" regardless of version, for example:

GSF Library files: C:\Projects\gsf \Build \Source etc...

Boost Library files: C:\Projects\boost \boost \libs etc...

If you have an existing Boost installation you can simply create a symbolic link to the folder, e.g., mklink /D C:\Projects\boost C:\boost_1_66_0

Alternately you can adjust the additional include directories to your own Boost installation location for each of the build configurations. The code has been tested with version v1.66 of Boost.

Note that you will need to compile Boost in order to execute the sample applications found in:

...\gsf\Source\Applications\TimeSeries Platform Library\Samples

The GSF Time Series Platform Library uses zlib features of Boost, as a result compiling boost requires zlib source code that can be downloaded separately: https://zlib.net/

After unzipping the zlib source code, set the following Boost compile script environmental variables to the root of the zlib source code path, e.g.:

set ZLIB_SOURCE="C:\zlib-1.2.11"
set ZLIB_INCLUDE="C:\zlib-1.2.11"

Once environmental variables are set for zlib paths, compile Boost as normal.

---

Compiling in Linux

The following information is intended to help developers build the GSF Time Series Platform Library on Linux platforms. Similar instructions may apply to other platforms.

1. Dependencies

---

The Time Series Platform Library depends on the following libraries in order to build. Earlier versions of the libraries listed may not work properly.

CMake v2.8 (http://www.cmake.org/)

GNU Make (http://www.gnu.org/software/make/)

zlib Library, e.g.: sudo apt install zlib1g-dev

bzip2 Library, e.g.: sudo apt install libbz2-dev

Boost C++ Libraries v1.66.0 (http://www.boost.org/) - Boost.Asio - Boost.Bind - Boost.Iostreams - Boost.System - Boost.Thread - Boost.Uuid

Boost will need to be compiled: https://www.boost.org/doc/libs/1_66_0/more/getting_started/unix-variants.html

For Ubuntu, here are some common steps (don't type $; that represents the shell's prompt):

$ sudo apt update $ sudo apt install build-essential $ sudo apt install cmake

$ sudo apt install zlib1g-dev $ sudo apt install libbz2-dev

$ sudo mkdir /usr/local/boost_1_66_0 $ cd /usr/local/ $ wget https://dl.bintray.com/boostorg/release/1.66.0/source/boost_1_66_0.tar.bz2 $ sudo tar -xvjf boost_1_66_0.tar.bz2

Start a new terminal session before building Boost:

$ cd /usr/local/boost_1_66_0 $ sudo ./bootstrap.sh $ sudo ./b2 install

It may be necessary to add /usr/local/lib, the default path for boost libraries, to the system library path before running any samples:

$ sudo ldconfig /usr/local/lib

2. Configuration

---

From the command terminal, enter the source directory containing this README.txt file and type the following command.

cmake .

Alternatively, you can create a build directory separate from the source code you downloaded. Enter the build directory you created and type the following command.

cmake path/to/source

Using the CMake GUI, you can modify configuration options, such as building as a shared library or changing the installation directory.

To make a debug build, use the following:

cmake -DCMAKE_BUILD_TYPE=Debug

3. Build

---

At the top level of the build directory, type the following command.

make

In addition to the library itself, there are sample applications which demonstrate the proper use of the Time Series Platform Library API. To build all samples, type the following command.

make samples

Individual sample applications can be built as follows:

make SimpleSubscribe
make AdvancedSubscribe
make AverageFrequencyCalculator
make InstanceSubscribe
make InstancePublish
make FilterExpressionTests
make SimplePublish
make AdvancedPublish

4. Installation

---

At the top level of the build directory, type the following command.

make install

This will move the header files and the library file to the location specified during configuration. Header files go under the 'include/' subdirectory, and the library file goes under the 'lib/' subdirectory.