This is the Fledge project.
Fledge is an open source platform for the Internet of Things, and an essential component in Fog Computing. It uses a modular microservices architecture including sensor data collection, storage, processing and forwarding to historians, Enterprise systems and Cloud-based services. Fledge can run in highly available, stand alone, unattended environments that assume unreliable network connectivity.
Fledge also provides a means of buffering data coming from sensors and forwarding that data onto high level storage systems. It assumes the underlying network layer is not always connected or may not be reliable. Data from sensors may be stored within Fledge for a number of days before being purged from the Fledge storage. During this time it may be sent to one or more historians and also accessed via a REST API for use by local analytical applications.
Fledge has been designed to run in a Linux environment and makes use of Linux services.
Fledge is built using a microservices architecture for major component areas, these services consist of:
- a Core service responsible for the management of the other services, the external REST API's, scheduling and monitoring of activities.
- a South service responsible for the communication between Fledge and the sensors/actuators.
- a Storage service responsible for the persistence of configuration and metrics and the buffering of sensor data.
This core set of services may also be extended using optional services that are available within their own repositories, for example a notification service and a control dispatcher service.
Fledge makes extensive use of plugin components in order to increase the flexibility of the implementation:
- South plugins are used to allow for the easy expansion of Fledge to deal with new South devices and South device connection buses.
- North plugins are used to allow for connection to different historians
- Datastore plugins are used to allow Fledge to use different storage mechanisms for persisting meta data and the sensor data
The South and North plugins are stored in separate source code repositories which are named in the pattern fledge-south-<device> fledge-north-<service>. A complete list of plugins can be found in the readthedocs documentation for the project. See Fledge Plugins.
The optional services, for example the notification service also make use of plugins to extend the capabilities of those services.
The other paradigm that is used extensively within Fledge is the idea of scheduling processes to perform specific operations. The Fledge core contains a scheduler which can execute processes based on time schedules or triggered by events. This is used to start processes when an event occurs, such as Fledge starting, or based on a time trigger.
Scheduled processes are used to send data from Fledge to the historian, to purge data from the Fledge data buffer, to gather statistics for historical analysis and perform backups of the Fledge environment.
Pre-built packages for Fledge are available, see Installing Fledge for details of how to use these.
See also Building Fledge in the online documentation.
Fledge is currently based on C/C++ and Python code. The packages needed to build and run Fledge may be installed by running the script requirements.sh
Fledge can be built or installed in one of the following Linux distributions :
- Ubuntu Ubuntu 18.04 and Ubuntu 20.04
- Raspbian Stretch, Buster and Bullseye
- Coral Mendel
The prerequisites required to build Fledge can be installed on any of the supported platforms by running the requirements.sh script from this directory.
sudo ./requirements.shNote
This script will use the platforms package management software, such as apt on Ubuntu systems, to install the required packages. This must be done as the root user and hence the need to run requirements.sh using the sudo command
To build Fledge run the command make in the top level directory. This will compile all the components that need to be compiled and will also create a runable structure of the Python code components of Fledge.
NOTE:
- The GCC compiler version 5.4 available in Ubuntu 16.04 LTS raises warnings. This is a known bug of the compiler and it can be ignored.
- openssl toolkit is a requirement if we want to use https based REST client and certificate based authentication.
Once the make has completed you can decide to test Fledge from your development environment or you can install it.
you can test Fledge directly from your Development Environment. All you need to do is to set one environment variable to be able to run Fledge from the development tree.
export FLEDGE_ROOT=<basedir>/Fledge
Where basedir is the base directory into which you cloned the Fledge repository.
Finally, start the Fledge core daemon:
$FLEDGE_ROOT/scripts/fledge start
Create an installation by executing make install, then set the FLEDGE_ROOT environment variable specifying the installation path. By default the installation will be placed in /usr/local/fledge. You may need to execute sudo make install to install Fledge where the current user does not have permissions:
sudo make install export FLEDGE_ROOT=/usr/local/fledge
The destination may be overriden by setting the variable DESTDIR in the make command line, to a location in which you wish to install Fledge. For example, to install Fledge in the /opt directory use the command:
sudo make install DESTDIR=/opt export FLEDGE_ROOT=/opt/usr/local/fledge
Fledge supports the Kerberos authentication starting from the version 1.7.1 and so the related packages are installed by the script requirements.sh. The krb5-user package prompt a question during the installation process asking for the KDC definition, the packages are installed setting the environment DEBIAN_FRONTEND to avoid this interaction:
# for Kerberos authentication, avoid interactive questions DEBIAN_FRONTEND=noninteractive apt install -yq krb5-user apt install -y libcurl4-openssl-dev
The upgrade of the Fledge package should follow the same philosophy, it should be done executing the command:
sudo DEBIAN_FRONTEND=noninteractive apt -y upgrade
before the upgrade of Fledge, SETENV: should be set/added in /etc/sudoers.d/fledge to allow sudo to support the handling of the environment variables, a sample of the file:
%sudo ALL=(ALL) NOPASSWD:SETENV: /usr/bin/apt -y update, /usr/bin/apt-get -y install fledge, /usr/bin/apt -y install /usr/local/fledge/data/plugins/fledge*.deb, /usr/bin/apt list, /usr/bin/apt -y install fledge*, /usr/bin/apt -y upgrade
Fledge is now ready to start. Use the command:
$FLEDGE_ROOT/bin/fledge start
To check if Fledge is running, use the command:
$FLEDGE_ROOT/bin/fledge status
The command returns the status of Fledge on the machine it has been executed.
This version of Fledge relies on SQLite to run. SQLite is embedded into the Storage service, but you may want to use PostgreSQL as a buffer and metadata storage (refer to the documentation on ReadTheDocs for more info. With a version of PostgreSQL installed via apt-get first you need to create a new database user with:
sudo -u postgres createuser -d <user>
where user is the name of the Linux user that will run Fledge. The Fledge database user must have createdb privileges (i.e. the -d argument).
The execution of the sudo make install immediately after git clone will create a data/etc directory owned by the root user, it should be owned by the user that will run Fledge, to fix it:
chown -R <user>:<user> $FLEDGE_ROOT/data
where user is the name of the Linux user that will run Fledge.
