Framework for Large Streaming Data Analysis

Since existing DoD ISR systems are of a typical stovepipe design, they are slow in speed and it is difficult to adapt them to changing information needs. Recent advancements in multi-media technologies add additional challenges to digesting audio and video stream data for the ISR missions. An increasing number of algorithms and data format conversion routines developed are available, but a clear bottleneck in the ISR community lies in the lack of a middleware framework that can horizontally and conveniently integrate heterogeneous computational resources. We aim to develop such a middleware framework, called INTERLINK, that is scalable and flexible in ingesting, analyzing and distributing streaming audio and video data for the DoD"s various needs. INTERLINK can bring a sweeping change to the way future ISR workflows are used. Web service providers will be able to plug-in new algorithms and data sources to the middleware framework in a matter of minutes through a publication and subscription model. The visual workflow interface will allow ISR professionals with limited computing expertise to easily construct and share workflows with others. We will not only develop engineering designs for INTERLINK but also a series of use case scenarios to demonstrate the power of the newly proposed middleware framework.