A real-time stream processing platform
The emerging Internet of Things is transforming the world into a giant source of live data streams. These streams range from slowly-trickling sensor samples to bandwidth-hungry video streams. They carry valuable data that – when processed in a continuous and timely manner – hold the potential to significantly improve people’s lives. For example, continuous streams of data can help us to make better-informed decisions, give us feedback for improved well-being, and connect homes and industries to create new levels of life automation. In addition to this, these devices will interact with each other and form control loops - and such control interactions require instant analysis and actions.
To execute on this vision - making individualized views of these data available, and triggering personalized actions and feedback - a new breed of software platforms is required – platforms that provide open, scalable, efficient, and programmable support for streaming data and fast-changing user contexts. Such open and programmable access to live streams will not only disrupt the way people consume and share streaming data – much like the original World Wide Web disrupted the way people consume and share predominantly stored data. It can be the seed of a much broader industry disruption – redefining how companies access, share & transform information streams.
Fragmentation in the IoT industry, rooted in disparate devices and applications built on proprietary protocols, can stifle innovation. This complex ecosystem makes it harder for application developers to innovate and create new applications cost effectively. With the rapid development of opportunities in the Internet of Things (IoT) marketplace, organizations are challenged in developing vertical business-specific solutions while ensuring maximum reusability across their organization. There is a need for software platforms that provide easy customization and onboarding of domain specific knowledge, algorithms and protocols.
Key features
A large-scale, geographically distributed, stream processing platform which can ingest, process and deliver large numbers of data and media streams in real-time between geographically distributed sources and sinks. Flexible chaining of pluggable analysis modules is possible, and new applications are authored centrally and distributed deployed on cloud, edge or device.
Applications on top of the World Wide Streams platform consist of one or more continuous queries or flows. Flows are authored in a novel Bell Labs language that is called XStream. XStream’s built-in operators cover most of the common stream processing operators (e.g. map, alter, join, partition, etc). Additionally, it supports the declaration of external operators which are not implemented in XStream itself, but whose inputs and outputs can be directly wired up with other XStream operators. XStream has been developed as a library and embedded in TypeScript, a JavaScript dialect with optional static typing.
Showcase and demonstrations
The connected car is going to play a fundamental role in the foreseeable Internet of Things. The connectivity aspect in combination with the available data (e.g. from GPS, on-board diagnostics, road sensors) and video (e.g. from dashcams and traffic cameras) streams enable a range of new applications. Five connected cars applications have been developed on top of World Wide Streams. Anomalous Driving is a scenario that detects cars that are rapidly accelerating or braking. Potential users of these scenarios can be fleet managers and parents. The Congested Areas application detects clusters of abnormal deceleration behavior, indicating rapid slow-down of multiple cars within a short time frame and geographic perimeter. Potential users of the events can be other drivers in the vicinity.
Remote Video Monitoring enables traffic officers to use a collection of the dashcam feeds within an area of interest (e.g. accident scene or congested square/crossing) to obtain insights in real time. Amber Alert allows a privileged user (e.g., a police department administrator) to initiate a simultaneous and large scale search for a particular license plate number (e.g., of a stolen car) within a certain geo-fence (e.g., a city district) using the available dashcams. It uses the dashcams to do license plate recognition, potentially leveraging the processing power available in the car. Sightings are reported on the map. Users can navigate to the car that did the sighting and see its live dashcam stream for immediate context. Aggregate Statistics provides city managers with real-time distribution of various events of interest such as speed, fuel consumption, and CO2 emission.
Get started now
Apply for an account to get access to our manual and stream away!
Please complete the form below.
The form is loading, please wait...
Thank you. We have received your inquiry. Please continue browsing.