Enabling Industrial Data Space Architecture for Seaport Scenario
Enabling Industrial Data Space Architecture for Seaport Scenario David Sarabia-Jácome, Ignacio Lacalle, Carlos E. Palau, Manuel Esteve Universitat Politècnica de València (UPV) Valencia, España
Presentation Outline • Introduction • Background • Seaport Data Space Architecture • Results • Conclusions • Future Research Directions
Introduction • Seaports are a transport and logistics infrastructure that have a significant economic impact on world trade. • The expected increase in port operations will make it difficult to manage the port’s resources and comply with the daily operational load requirements, which will result in unproductive operations. • Industry 4. 0 as a solution to promote seaport process efficiency. • Automation of port operations and equipment employing Cyber. Physical Systems (CPS). • Imminent heterogeneity of Io. T platforms, suppliers, technologies as a limitation.
Motivation • Data ownership issue. They are afraid of losing control over the data shared. • Industrial Data Space (IDS) standard initiative proposes a reference architecture model to cover the data ownership issue (Data sovereignty). • IDSA motivates the interaction among researchers and business to actively develop the standard by mean of use cases implementations. Seaport case were not considered yet.
Propose • Designing an architecture to sharing data between seaport stakeholders based on IDS architecture. • Implementing a tesbed to evaluate the feasibility of employing IDS architecture to solve the interoperability (data sovereignty) problem in seaport scenario. • Designing and implement a Big Data Architecture to exploit the data shared in the seaport data space. • Developing KPIs employing the data shared to improve the seaport operations.
Background IDS Architecture • Five layers to describe the entities, functional roles and interaction: business, functional, information, process, system. • Business layer describe the roles key roles (Data Provider, Data Consumer, Data Owner, Data User), and entities (IDS Connector, App Store, Broker, Identity Provider) App Store Broker App Metadata Data Provider Connector Data exchange Identity Provider Data Owner Data Consumer Connector
Background IDS Architecture • IDS Connector: key entity capable of interconnecting among them and exchange data. Use a push/pull mechanism. Unique identification and certificated. Execute apps in the client site. To ensure data sovereignty use PEP proxy. • IDS Broker: Managing information of data source in the space. Stores metadata, semantics information, pricing or usage policies. • Identity Provider: Identity manager service. Store information about IDS connectors to ensure trust and security. • App Store: repository of service and applications to process data in the IDS connector.
Seaport Data Space Architecture Port Authority Io. T Platform IDS Connector Container Terminal Sharing Data IDS Connector Big Data Analytics
Big Data Architecture Batch Processing Statistical Analysis Pre-Processing Serving Long Term Store Real time Processing Integration IDS Connector
Testbed Implementation XACML IDS CONNECTOR Proxy (Wilma) PEP NGSI Io. T Agent Weather Id. M (Keyrock) FIWARE Context Broker (Orion) NGSI System Adapter AIS Token BIG DATA ANALYTICS IDS CONNECTOR NGSI FIWARE Context Broker (Orion) Proxy (Wilma) Apache Flink NGSI PEP NGSI System Adapter CSV Cosmos (HDFS) Apache Spark/Hadoop Apache Kaf. Ka Token PAP/PDP Auth. ZForce Cygnus Id. M (Keyrock)
Testbed Implementation (IDS Connector) Id. M (Keyrock) Token • PAP/PDP Auth. ZForce XACML • IDS CONNECTOR Proxy (Wilma) PEP NGSI Io. T Agent Weather FIWARE Context Broker (Orion) NGSI System Adapter AIS NGSI • • • FIWARE-based connector employing Orion Context Broker as its core. Io. T agent (MQTT) to connect to weather sensors and a System adapter to connect to AIS. HTTPS NGSI API to ensure secure communication. Wilma Proxy ensures the usage policies and usage enforcement. Data is kept on the Mongo DB attached to Orion Context Broker.
Testbed Implementation (IDS Connector) Id. M (Keyrock) Token • PAP/PDP Auth. ZForce XACML • IDS CONNECTOR Proxy (Wilma) PEP NGSI Io. T Agent Weather FIWARE Context Broker (Orion) NGSI System Adapter AIS NGSI • • • FIWARE-based connector employing Orion Context Broker as its core. Io. T agent (MQTT) to connect to weather sensors and a System adapter to connect to AIS. HTTPS NGSI API to ensure secure communication. Wilma Proxy ensures the usage policies and usage enforcement. Data is kept on the Mongo DB attached to Orion Context Broker.
Testbed Implementation (IDS Connector) • A push mode Context broker federation. • The notify. Context are sent to the other IDS connector using NGSI API (HTTPS).
Testbed Implementation (Big Data Architecture) BIG DATA ANALYTICS NGSI Cygnus Cosmos (HDFS) Apache Spark/Hadoop Apache Kaf. Ka Apache Flink • Context Broker only stores the last value. • Cygnus GE collects data from Orion Context broker and move to the long-term storage. • HDFS is a high availability and scalability data storage. • A Parquet file is saved in HDFS for each entity subscribed and updated for each NGSI context notification received. • Big Data Analysis Cosmos GE is used to provide batch (Apache Spark) and stream processing (Apache Flink). • Spark. SQL high level language to perform a descriptive analysis.
RESULTS (Valencia Port Use Case) • Descriptive data analysis carried out with three datasets (Weather, AIS, Terminal Operation) Dataset Weather AIS Terminal Operation Size 28 MB 10 GB 52 MB Period 2014/01/01 – 2017/08/31 2016/01/01 – 2016/03/31 2014/01/01 – 2017/08/31 • Development of two KPIs for seaport operations: • Average time spent per vessel in terminal. • Terminal occupancy.
RESULTS (KPI - 1) • Trend, season, and residual data characteristics are evaluated from the vessels load/unload dataset.
RESULTS (KPI - 2) • Months and days of the week on which the terminal is the most occupancy
RESULTS (KPI - 2)
RESULTS (KPIs) Parameters KPI 1 [hour] KPI 2 [Vessels] Average per week 23, 3065 28 Max per week 64, 2782 10 Min per week 1, 8544 52
RESULTS (Web UI)
Conclusions • A seaport data space is presented employing open source Io. T platform (FIWARE) following the IDS reference model architecture. • The implementation of the seaport data space for Valencia port demonstrates the feasibility of the IDS in seaports scenarios. • The design and implementation of a Big Data Architecture to exploit the shared data in the seaport data space takes advantage of it and ensure the data ownership. • Two KPIs were developed to validate the Big Data Architecture and the sharing between IDS connectors processes.
Future Research Directions • Other stakeholders (hauliers company, terminals passengers, among others) will be add to the seaport data space. • The rest of the IDS entities that are not considered in this paper will be implemented. • More KPIs will be developed to enrich the seaport operation dashboard.
- Slides: 22