A trip through the Ozone
In the excitement over big data, much of the discussion has focused on analytics – using data “facts” to derive interesting conclusions about specific areas of interest. The attention has been on sophisticated algorithms and the technologies that implement them on the back end.
As complex as these algorithms and technologies can be, they only tell half the story. Eventually humans need to consume the results of analytics. This means visualizations on the front end are just as essential to reaping the benefits of big data. As we’ve seen with commercial products like Tableau and open-source products and tools like Gephi and D3, interest in cool visualizations to supplement cool analytics continues to grow.
Long a proponent of big data analytics, the intelligence community also recognizes the role visualizations play in helping analysts make the critical decisions necessary to safeguard national security and to empower the warfighter. To that end, the community has worked with private contractors and the Government Open Source Software Board to build and open source the Ozone Widget Framework.
To understand OWF, let’s first explore the concept of a widget. Imagine one contractor builds a map to render geospatial data in a browser, another builds a timeline to render temporal data in a browser, and a third builds a viewer to render tabular data in a browser. OWF can unite all three disparate Web applications within a single dashboard by providing a small window displaying each within the browser simultaneously. Each of these windows is a widget. Here is an image from the OWF website that explains this idea:
Developers are empowered to build applications without any knowledge they will ultimately be “widgetized” in OWF. That’s a configuration detail for an integrator later.
It gets better. Despite the independent origins of these applications, OWF uses the publish-subscribe pattern (one of the Enterprise Integration Patterns I described in a previous column ) to endow widgets with the ability to communicate with one another. For example, the map widget can subscribe to messages broadcast by the table widget and re-render itself with new geographical data accordingly. This interactivity among once-isolated applications lies at the core of what makes OWF so powerful.
Taken together, the widgets on the OWF dashboard can provide different views into the same data and allow analysts to draw meaningful conclusions that might otherwise be impossible.
OWF is essentially a Java Web application. More precisely, it is built in Grails, a Web framework based on the Groovy programming language that runs in the same environment as Java. This means OWF is deployed to Web servers like Tomcat and JBoss just as with traditional Java Web applications. OWF can tailor dashboards to user preferences stored in a relational database and to permissions stored in something like Lightweight Directory Access Protocol (LDAP), which are managed by Hibernate and Spring Security respectively. Both are popular frameworks in the Java open-source ecosystem.
According to Ozone website, OWF is supported on IE7+ and Firefox 3.6+.
Impact of OWF
Though initially developed by private contractors under the auspices of the National Security Agency, development of OWF now spans multiple agencies, and the broader intelligence community has taken notice of its immense potential. OWF is the front end for the analytics that drive the U.S. Army Intelligence & Security Command Futures effort. The Coast Guard and the Defense Information Systems Agency have also adopted OWF.
Aside from its value as an analytic tool, OWF carries broader significance. The 2012 National Defense Authorization Act required OWF to be open source, so the code is available on GitHub for anyone to fork under the Apache 2.0 license. It is a credit to the Defense Department that it has taken a position of leadership in bringing the vibrancy and innovation of the open-source community into the government space.
OWF continues to be a work in progress. While new features are developed, there is an effort to refactor the existing code base to remove the dependencies on Grails and Sencha Ext JS and to improve performance and scalability. OWF must improve its configurability and modularity so it works in different environments. Related projects like Ozone Marketplace and OWF Mobile need to evolve. Still, there is no denying the potential for OWF to prove a valuable tool for intelligence analysts and a significant victory for government open-source advocates.