Sample time-series application using car telemetry as the use case
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.

README.md

Racing Time Series

This application shows how to use MapR Converged Data Platform to capture data from Racing Cars for:

  • Real Time Analytics
  • Big Data Processing

The application is built using various components:

  • The Open Racing Car Simulator : used to generate metrics data such as speed, rpm, and lap during races
  • The car metrics are:
    • captured using a MapR Streams Producer
    • consumed by a consumer that push data in real time into a Web Browser.
    • another consumer is saving all the events into a MapR-DB JSON table for further analytic purpose

Prerequisites

  • MapR 5.1 or later Cluster with MapR-DB and Streams enabled, this could be the MapR 5.1 Sandbox
  • Vagrant, the demonstration is running in a Vagrant Virtual Machine
  • Git to clone the demonstration content

Installation

The "How to setup the demonstration on a stand alone Mac" will help you to configure the Virtual Machines and Network; you can adapt this to Windows and Linux easily.

1- Clone the project

$ git clone https://github.com/mapr-demos/racing-time-series.git

2- Create Folders and Streams

The demonstrations is using:

  • Topics located in the /apps/racing/stream stream
  • Tables located in the /apps/racing/db/telemetry folder

To create all these object run the ./demo/setup-stream.sh file on your MapR cluster.

Note: you can run this script any time to reset the entire environment.

3- Configure the project

As mentioned above the demonstration will be running in a Vagrant Virtual Machine, that will connect to your MapR Cluster. The first thing to do is to configure the project to match your environment.

Open the ./demo/config.conf file and edit the following properties:

  • MAPR_UID and MAPR_GID : the mapr user id and mapr group id used by your cluster to be sure the MapR client and cluster are using the same ids
  • CLUSTER_IP one of the nodes of your cluster
  • CLUSTER_NAME the name of your MapR cluster

Note: The cluster IP and name will be used to configure the MapR Client, you will be able to modify that later if necessary.

4- Initialise the Racing demonstration

You will now use Vagrant to create a new Virtual Machine, define by the ./demo/Vagrantfile. (You should not have to change the configuration). In a terminal do the following:

$ cd demo
$ vagrant up

This operation will create a new Ubuntu virtual machine and execute the demo/setup.sh script.

The setup.sh is executed automatically the first time, and it takes a while, as it will:

  • Update Ubuntu
  • Install additional package including MapR Client and MapR Stream client (mapr-kafka package)
  • Download TORCS source code, the race simulator
  • Apply a patch to TORCS to allow the demonstration to capture racing cars telemtry information
  • Build TORCS
  • Build the Racing Telemetry demonstration itself (Java projects located in ./demo/racing-telemetry-application)

5- Check the installation

You should be able to connect to the VM using the 'vagrant' user with password 'vagrant', you can also connect to the VM using ssh using the command: vagrant ssh, when you are in the ./demo/ folder.

In a terminal, in the Ubuntu VM, run the following command:

  • hadoop fs -ls / : you should see the list of folders of you MapR cluster

If you encounter errors it is most of the time due to some network issues, you must be sure that your Virtual Machine can access the MapR cluster.

Running the Demonstration

  1. If the Ubuntu VM is not running, using a terminal go to ./demo/ and run vagrant up

  2. Go to the Ubuntu desktop, if the screen is locked, log with the user vagrant password vagrant.

  3. On the desktop click on the "Streams Demo" icon, this will start multiple windows.

  4. After few seconds you should see the TORCS simulator and a Web Browser

  5. In the TORCS Simulator, click on Race, Quick Race, New Race

  6. This should start a race with 3 cars, and send data to your Web Browser, where you can select various metrics.

Note:

  • The current Web site only print the data of the current race
  • You can customise the race in TORCS but only the Inferno cars are emitting data to Streams.
  • The demo folder of the project is automatically loaded as a shared folder into the Vagrant virtual machine and accessible at /vagrant

Data Flow - How this works?

The application is made of 3 major components:

  • The TORCS, that as been updated to emit Inferno cars metrics into a file. /usr/local/share/games/torcs/telemetry/Inferno.dat on the Vagrant VM.
  • A Kafka Producer, that reads the file to post metrics on the /apps/racing/stream:events topic.
  • The demonstration uses several Consumers to split information by cars, but also send data in real time to the Web interface using Web socket.
    • A generic Consumer that split the events to many topics, one generic and one by car
    • CarStreamConsumer to save data in one table by car (/apps/racing/db/telemety/car1, ..., car10)
    • A Web application with a EventSourceServlet that use the LiveConsumer to send event to the browser using Websocket. The LiveConsumer subscribe to individual car topics.

The most important message type looks like:

{
	"sensors" : 
	{
		"Speed":51.13699,
		"Distance":4051.190186,
		"RPM":706.666504
	}
	,
	"racetime":73.738,
	"timestamp":1458127205
}
  • Speed in km/h
  • Distance in km
  • RPM
  • Racetime: time in the race
  • timestamp: start time of the race