SC4 Pilot Online Hangout
Luigi Selmi - FhG27 June 2016
Pilot Scenario
Multisource data collection for the provision of
accurate info-mobility and advanced transport
planning services in Thessaloniki
◎FCD Taxi
◎BT P2P Detectors
Present FCD Pipeline
Storage and processing in MSSQL Server
Aims of the SC4 Pilot
A new system with Big Data characteristics
◎Volume (Scalability and Fault-tolerance)
◎Velocity (Low latency)
◎Variety (Formats, Unstructured, Structured)
◎Easy Deployment
Reference Architecture
SC4 Pilot Architecture
Semantic Lifting
To improve the quality of the data and make a
better use of it, the semantics of the data will
be made explicit at the ingestion phase using
standard ontologies or creating new ones.
SC4 Pilot Components
◎Data Ingestion: Kafka Producer (Flink Jobs)
◎Data Integration: Kafka
◎Data Processing: Flink + Rserve + PostGis
◎Data Storage: Elasticsearch
◎Data Presentation: Kibana
◎Pilot Pipeline
Kafka & Flink
SC4 Pilot Github Repos
5 Projects on the BDE Github Repository
◎Pilot-sc4-kafka-producer
◎Pilot-sc4-docker-r
◎Pilot-sc4-mapmatcher (for testing)
◎Pilot-sc4-flink-kafka-consumer
◎Pilot-sc4-pipeline (docker-compose)
BDE Components
The BDE platform provides
◎Docker images for all the components
◎A tool for integration and initialization based
on Docker Swarm
◎Integrated UI
3 Nodes Configuration
Current and Future Work
1. Set up a Docker Swarm for the pilot in 3
nodes (real and VMs) using the dockerized
version of all the components with Init
Daemon and BDE Integrated Dashboard
2. Test the pilot swarm for fault-tolerance and
scalability

SC4 Hangout - Luigi Selmi, Transport pilot architecture

  • 1.
    SC4 Pilot OnlineHangout Luigi Selmi - FhG27 June 2016
  • 2.
    Pilot Scenario Multisource datacollection for the provision of accurate info-mobility and advanced transport planning services in Thessaloniki ◎FCD Taxi ◎BT P2P Detectors
  • 3.
    Present FCD Pipeline Storageand processing in MSSQL Server
  • 4.
    Aims of theSC4 Pilot A new system with Big Data characteristics ◎Volume (Scalability and Fault-tolerance) ◎Velocity (Low latency) ◎Variety (Formats, Unstructured, Structured) ◎Easy Deployment
  • 5.
  • 6.
  • 7.
    Semantic Lifting To improvethe quality of the data and make a better use of it, the semantics of the data will be made explicit at the ingestion phase using standard ontologies or creating new ones.
  • 8.
    SC4 Pilot Components ◎DataIngestion: Kafka Producer (Flink Jobs) ◎Data Integration: Kafka ◎Data Processing: Flink + Rserve + PostGis ◎Data Storage: Elasticsearch ◎Data Presentation: Kibana ◎Pilot Pipeline
  • 9.
  • 10.
    SC4 Pilot GithubRepos 5 Projects on the BDE Github Repository ◎Pilot-sc4-kafka-producer ◎Pilot-sc4-docker-r ◎Pilot-sc4-mapmatcher (for testing) ◎Pilot-sc4-flink-kafka-consumer ◎Pilot-sc4-pipeline (docker-compose)
  • 11.
    BDE Components The BDEplatform provides ◎Docker images for all the components ◎A tool for integration and initialization based on Docker Swarm ◎Integrated UI
  • 12.
  • 13.
    Current and FutureWork 1. Set up a Docker Swarm for the pilot in 3 nodes (real and VMs) using the dockerized version of all the components with Init Daemon and BDE Integrated Dashboard 2. Test the pilot swarm for fault-tolerance and scalability