Cloud4All — Context-Aware-Server

Description

The Context Aware Server (CAS) developed in NodeJs provides a platform that collects, stores and process the data from sensors and when some triggers are fired it sends the processed data to a client in json format. Configuration, Sensors and Users are stored in MongoDb and new data is cached in Redis and it can be eventually saved to MongoDb depending on configuration.

License

This project has been developed by Barcelona Digital Health department. Device reporter is shared under New BSD license. This project folder includes a license file. You can share and distribute this project and built software using New BSD license. Please, send any feedback to http://www.cloud4all.info

Installation

Download de zip file and uncompress it.

cd Cloud4All---Context-Aware-Server-master
npm install

It will install all needed packages that can be found in package.json file. Aftwards, you need to install Redis and Mongo databases. In ubuntu/debian

$ apt-get install redis-server

Quick Start & Examples

The best way to get started with the CAS is running it using node

node app.js

Then you can then send a POST request to the CAS to add a new sensor. For example, you can send the sample in test/data folder

curl -H "Content-Type: application/json" -X POST --data @test/data/new-device-sample.json http://localhost:8888/devices

Afterwards you can send a POST request to the CAS to add data to the new sensor.

curl -H "Content-Type: application/json" -X POST --data @test/data/sensor-sample-data.json http://localhost:8888/sensors/{uuid}/data

And get all the data stored in CAS from that sensor

curl http://localhost:8888/sensors/{uuid}/data

Or also ask for data from :start time and :end time in ISO-8601 format using following request

curl http://localhost:8888/sensors/{uuid}/data/start/2013-04-22T00:35:43.12Z/end/2013-04-22T01:15:43.28Z

Setup

CAS can be configured using a Restful API. To setup a new user or sensor configuration, send a POST request to /configs/ where _ref indicates the :id of the user or sensor.

{
	"_ref": "517686388661d24a16000006",
	"config": {
        "triggers": {
            "onNewData": {
                "onConfig": "getNewData",
                "onData": "getSumData",
                "onAggregated" : "threshold",
                "onTriggered": "getNearUsers",
                "onNearby": "sendData",
                "onSent": "storeData",
                "onStored": "ack",
                "diffRadius": "10",
                "threshold": "5",
                "maxDistance": 1
            },
            "scheduling": {
                "data": "mean",
                "time": "60"
            }
        },
        "receiver": {
            "host": "localhost",
            "port": "8889",
            "path": "/receiver"
        }
    }
}

Updating the configuration is possible sending a POST request to /configs/:id where id is the id of the sensor or user. You can also get a specific config system. Use GET to retrieve the value from key config parameter from id configuration:

GET /configs/:id

Finally, you can delete users, configs and devices sending a delete request to specific 🆔

DEL /configs/:id
DEL /devices/:id
DEL /users/:id

Triggering System

When new data arrives to the Context awareness server, it fires a "onNewData" event. Afterwards, a listener gets the sensor configuration from the mongo database. Then the trigger system will emit specific events depending on the configuration. These events will collect, process the data and send the data to the receiver if the trigger (for example the threshold) is surpassed. If nearby options are activated, it will search the near users and it can afterwards send the user's profile complemented with the context data.

For example, using the sensor:1 configuration above, when new data arrives, the trigger system will collect only new data "getNewData" from the sensor, will sum "getSumData" it and then uses a simple threshold "threshold" method to know if this sum is above the required value "5", if it is such the case, it will send then search the near users "getNearUsers" and will send the new data "sendData" to the "localhost" receiver and also store it to MongoDb "storeData".

Aggregation Methods

sum = sum of all values
mean = mean of all values
max = max of all values
min = min of all values
last = last of new values

Data Query methods

all = all stored data
new = only new data

Trigger methods

1. threshold = triggered when the value is above a threshold. threshold value must be defined. "threshold": "{number}"
2. diffRadius = triggered when there is a the last new value y and the last sent value x do comply the inequattion |y - x| > {number} / 100 * x where {number} is the defined value in configuration "diffRadius": "{number}". Default 10%.

User Filter methods

1. getNearUsers = once triggered, check which users are in nearby. $MaxDistance can be configured.

Send methods

1. SendData: once system is triggered, it sends the data received via json to a server (configured in settings).
2. SendProfile: once system is triggered, it sends users filetered from User Filter Methods. only useful when Filter methods are applied.

Running Tests

Test are implemented using mocha. To run the test suite first invoke the following command within the repo, installing the development dependencies:

npm install
make test

An Introduction Example

First of all, you have to start from scratch. Remove from mongodb the devdb (and also testdb). Then:

• Run the mongo server (if it is not running already):

mongod

• Run the node app. We have to be in root directory (where app.js is):

node app.js

• Create a new device using the sample data in test

curl -H "Content-Type: application/json" -X POST --data @test/data/new-device-sample.json http://127.0.0.1:8888/devices

• Add real time samples to sensor: The terminal outputs the _id of the sensors. Use one of them in the following command

node tools/sensor-mockup.js --id={{_id selected}}

• Create a trigger: Go to the web and search for the sensor that is receiving data (in http://localhost:8888/). Select the sensor that is receiving data (you can see it because its Value is changing), in the next page create a trigger clicking on [Add]. For example, add a "threshold greater than 5", and Submit.

• Connect the socket.io to stream new triggered data from specific location. Check for the gps location of the device created before (in our example is "gps": 2.194909, 41.402455. In the browser, go to the following url:

  http://localhost:8888/context/location/?gps=2.194909,41.402455

If all is going smothly, you will see that fired triggers with data are sent to this specific client using web sockets.

Contributors

Guillem Serra from Barcelona Digital

Funding Acknowledgement

The research leading to these results has received funding from the European Union's Seventh Framework Programme (FP7/2007-2013) under grant agreement No.289016 (Cloud4all).