Lattice Monitoring Framework

This project contains both the source code and some pre-compiled binaries of the Lattice Monitoring Framework.

The main components of the framework are:

• The Controller
• Data Sources
• Data Consumers

The Intelligent MOnitoring Subsystem (IMoS) of the 5GEx MdO interacts with the Lattice Controller using a REST API (which represents an implementation of the I3-Mon interface).

In general, we can assume that a single Controller instance is up and running in a given 5GEx technological domain. However, alternative monitoring topologies might be used.

Build

The provided build.xml (under source/) can be used both for compiling the source code and generating the deployable jar files.

$ cd source/
$ ant dist

The above command generates two different jar binary files in the jars directory:

• monitoring-bin-controller.jar containing all the classes and dependencies related to the controller.
• monitoring-bin-core.jar containing a subset of classes and dependencies that can be used for instantiating Data Sources and Data Consumers.

and also a jar containing the source code

• monitoring-src.jar

Installation

As soon as the Build process is completed, the controller can be started as follows:

$ cd jars/
$ java -cp monitoring-bin-controller.jar eu.fivegex.monitoring.control.controller.Controller controller.properties

The controller.properties file contains the configuration settings for the controller (an example is reported under conf/)

Configuration

info.localport = 6699

is the local port used by the Controller when connecting to the Information Plane. Other Lattice entities (e.g., Data Sources) will remotely connect to this port once started.

restconsole.localport = 6666

is the port where the controller will listen for HTTP control requests coming from the orchestration layer (i.e., IMoS).

probes.package = eu.fivegex.appl.probes
probes.suffix = Probe

When generating the probe catalogue, the Controller will look for all the class files whose package name matches eu.fivegex.appl.probes and whose name terminates with the Probe suffix (e.g., eu.fivegex.appl.probes.docker.DockerProbe.class).

deployment.enabled = true

Can be set either to true or false and enables/disables respectively the automated Data Sources deployment functionality to a remote host (current implementation is based on SSH with public key authentication)

deployment.localJarPath = /Users/lattice
deployment.jarFileName = monitoring-bin-core.jar
deployment.remoteJarPath = /tmp
deployment.ds.className = eu.fivegex.monitoring.appl.datasources.ZMQDataSourceDaemon
deployment.dc.className = eu.fivegex.monitoring.appl.dataconsumers.ZMQControllableDataConsumerDaemon

The above settings allow to specify (in order):

• the path where the jar (to be used for the Data Sources / Consumers automated remote deployment) is located
• the file name of the above jar file
• the path where the jar will be copied on the remote machine where the Data Source is being deployed
• the class name of the Data Source to be started (it must exist in the specified jar)
• the class name of the Data Consumer to be started (it must exist in the specified jar)

Testing

In order to verify that the Controller is up and running, a quick check consists in performing a GET request to /probe/catalogue/. The expected result is a response code 200 and Content-Type: application/json.

Usage for the vCDN experiment

Lattice is the monitoring backend that will enable collecting measurements from the VNFs deployed via the 5GEx Orchestration System. It can be used e.g., to retrieve measurements from Docker containers and / or Openstack VMs. When Lattice is used with Docker, the system running the Docker Engine has to be configured to expose the REST API on port 4243 (see Docker documentation).