Scenario: Multiple Sites

Scenario

An institution with multiple field sites wants to use DataTurbine to reliably buffer and stream data to a central database. Several partner institutions are interested in accessing this data.

The System

Multiple field stations, all with internet access run complex instrument packages. The link between them and the central server is unreliable and prone to downtime.

Each site runs its own DataTurbine Server to buffer the data and mirrors that data to a central server also running Data Turbine. That server runs a suite of analysis applications and stores the data in a database.

Several partner institutions also utilize this data. They have their own deployments running and may likewise contain multiple sites. They run a multitude of sources to get data and use analysis routines that utilizes information from multiple institutions and sites. They store the data and the result of the analysis in their own database.

Each institution has users that visualize the data via MatLab, LabView, Website, RDV, Google Earth, and other applications

The Parts

  1. Field Sites
    1. Each with a Data Turbine Server (rbnb.jar)
    2. Each Running Multiple Sources
  2. Central Server
    1. DataTurbine Server (rbnb.jar)
    2. Database & DT2DB
    3. Other DataTurbine Applications
  3. Partner Institutions
    1. Each with a DataTurbine Server (rbnb.jar)
    2. Each receiving Data from Multiple Sources
    3. Each running DataTurbine Applications (sources/sinks)
  4. Clients
    1. Each Running Visualization Applications (sinks)

Procedure

  1. At Central Server
    1. Setup and run a DataTurbine Server (rbnb.jar)
    2. Setup DT2DB to save data streams to Database
    3. Setup any analysis applications that are desired
    4. Ask network administrator to open up port ~3333 on the server
  2. At Field Stations
    1. Setup and run DataTurbine Server (rbnb.jar)
    2. Start a series of sources to load data into the local DataTurbine instance
    3. Set a push mirror targeting the central server
  3. At Partner Institutions
    1. DataTurbine is already running
    2. Sources are already running
    3. Other applications are already running
    4. The mirror will be a pull mirror so port does not have to be open
    5. Partner institution sets up pull mirror targeting the central server of the home institution
  4. Clients
    1. Any sink can connect to the home institution server
    2. If ports are open on the partner institution servers, sinks can connect to those as well
    3. Sinks should not connect to the field sites. The ports at the field sites should be closed as public access can overload those machines

Multiple Sites

Open Source DataTurbine Initiative © 2016 Frontier Theme