Back to Top

D8.1 - First Report on Warsaw Use Case

The aim of this report is to provide overall description of the City of Warsaw prototype of VaVeL platform. The key aspects related to CoW use-cases are: (i) processing of live stream of tram data, (ii) processing of live stream of bus data and (iii) linking live streams with schedules and other static information. The report gives overview on the test beds and production CoW VaVeL platform. The test beds are installed in Orange Polska premise (’bajorko’ cluster) and Warsaw University of Technology, and are both based on Apache Hadoop ecosystem, mostly on customized Horton Data Platform installation. The production platform used within Warsaw pilot will be hosted at the City of Warsaw IT Department and will follow the same guidelines and design pattern on both test beds. The report provides deep technical description of component developed especially for CoW pilot, including, but not limited to:
  (a) Apache Flink-based components for live stream processing:
       (a) Integration of live streams with schedule data
       (b) Filtering of tram and bus data
       (c) Estimating delays in trams/buses
       (d) Estimating type of movement for trams/buses
  (b) Apache Flume based components devoted to data collection, especially Flume’s components acting as a gateways to external systems, data sources etc., including but not limited to:
       (a) Polling data source about live trams and buses data
       (b) Collecting ZTM (Warsaw’s Public Transport Authority) RSS channel news
       (c) The backend components for mobile application in Warsaw pilot
The technical description of the modules provides the guidelines on modules installation, running and detailed description of processed data. Moreover the expected inputs and outputs are included in order to simplify the verification of proper functioning of the modules in new environment. This directly addresses the VaVeL project interoperability goal which aims the platform to be transferable to large number of cities after the project is finished. The mobile application backend (MAB) is based on components out of standard Hadoop ecosystem, but its main goal is to support mobile application, eliminate unnecessary load on the platform and minimize network traffic. There is necessity of providing expected scalability and performance. MAB consumes processed data streams from the core of VaVeL platform and makes them available on a subscription-based mechanism to the mobile application users. The document was prepared as part of Task T8.1- Prototype Development.