390 likes | 406 Views
This overview discusses real-time shipping business goals and the high-level solution for accurate prediction and immediate detection of main engine failures and issues related to sensor data for enhanced customer satisfaction and safety.
E N D
BigDataStackData Services OverviewJune 26th 2019 Yosef Moatti – IBM Haifa Research Labs
Agenda • BigDataStack project Overview • Danaos Real-time Shipping • BigDataStack Data Services Overview Periodic Review Meeting
Real time shipping: business Goals Recap of the Business Goals: • Accurately predict specific main engine failures • Gets informed of “problems” ASAP • Broken sensors • Infringement of business rules (e.g., fuel consumption) • Build a solution which will also fit the coming Danaos data ingestion rates: Periodic Review Meeting
Real time shipping: business Goals Recap of the Business Goals: • Accurately predict specific main engine failures • Gets informed of “problems” ASAP • Broken sensors • Infringement of business rules (e.g., fuel consumption) • Build a solution which will also fit the coming Danaos data ingestion rates: Periodic Review Meeting
Real time shipping: business Goals Recap of the Business Goals: • Accurately predict specific main engine failures • Gets informed of “problems” ASAP • Broken sensors • Infringement of business rules (e.g., fuel consumption) • Build a solution which will also fit the coming Danaos data ingestion rates: Periodic Review Meeting
Real time shipping: business Goals Recap of the Business Goals: • Accurately predict specific main engine failures • Gets informed of “problems” ASAP • Broken sensors • Infringement of business rules (e.g., fuel consumption) • Build a solution which will also fit the coming Danaos data ingestion rates: Periodic Review Meeting
Real time shipping: business Goals Recap of the Business Goals: • Accurately predict specific main engine failures • Gets informed of “problems” ASAP • Broken sensors • Infringement of business rules (e.g., fuel consumption) • Build a solution which will also fit the coming Danaos data ingestion rates: Periodic Review Meeting
Real time shipping: business Goals Recap of the Business Goals: • Accurately predict specific main engine failures • Gets informed of “problems” ASAP • Broken sensors • Infringement of business rules (e.g., fuel consumption) • Build a solution which will also fit the coming Danaos data ingestion rates: Periodic Review Meeting
Real time shipping: business Goals Recap of the Business Goals: • Accurately predict specific main engine failures • Gets informed of “problems” ASAP • Broken sensors • Infringement of business rules (e.g., fuel consumption) • Build a solution which will also fit the coming Danaos data ingestion rates: Periodic Review Meeting
Constraints on the Solution • Alerts must be triggered on streaming data • Access to recent data should be highly performant • Data scientists wish not to mind about data location Periodic Review Meeting
High level view of the solution Vessels IoT Devices Gateway Edge Seamless component BigDataStack Data Center
High level view of the solution Streaming data handled within data center Streaming data handled at Edge Vessels IoT Devices CEP CEP Gateway Streaming Data Management Edge Seamless component BigDataStack Data Center
High level view of the solution Remote IBM COS Historic data Historic data Recent data Vessels IoT Devices LXS DB CEP CEP Gateway Edge Seamless component BigDataStack Data Center
Recent data in LXS can be accessed directly Remote IBM COS `1 Data service Data scientist Vessels IoT Devices LXS DB CEP CEP Gateway Edge Seamless component BigDataStack Data Center
Data Services which directly access LXS DB Remote IBM COS `1 Danaos Preventive Maintenance alg. Data Quality Assessment Vessels IoT Devices LXS DB CEP CEP Gateway Edge Seamless component Danaos Alarm detection daemon BigDataStack Data Center
Historic data in Object Store can be accessed directly Remote IBM COS Data service Data scientist Vessels IoT Devices LXS DB CEP CEP Gateway Edge Seamless component BigDataStack Data Center
Dataset can be accessed seamlessly Remote IBM COS Data service Data scientist Vessels IoT Devices LXS DB CEP CEP Gateway Edge Seamless component BigDataStack Data Center
Recent data in LXS can be accessed directly Remote IBM COS `1 Data service Data scientist Vessels IoT Devices LXS DB with Data Skipping Data Mover CEP CEP Gateway Edge Seamless component BigDataStack Data Center
Phase 1: processing at Edge Sensor data Periodic Review Meeting
Phase 1: processing at Edge CEP at Edge At Edge Sensor data Periodic Review Meeting
Phase 1: processing at Edge CEP at Edge At Edge CEP1 Sensor data Basic rules check Periodic Review Meeting
Phase 1: processing at Edge CEP at Edge At Edge CEP1 2018-10-23 08:00:00,stw,15.062018-10-23 06:36:00,stw,15.162018-10-23 15:56:00,stw,14.92018-10-23 15:57:00,stw,14.922018-10-24 08:30:00,stw,16.182018-12-21 18:42:00,stw,17.112018-12-21 18:43:00,stw,16.922018-12-21 18:45:00,stw,16.772018-10-23 16:01:00,stw,14.74 Sensor data Sensor data 2018-10-23 08:00:00,stw,15.06, 02018-10-23 06:36:00,stw,15.16, 02018-10-23 15:56:00,stw,14.9, 02018-10-23 15:57:00,stw,14.92, 12018-10-24 08:30:00,stw,16.18. 02018-12-21 18:42:00,stw,17.11, 02018-12-21 18:43:00,stw,16.92, 02018-12-21 18:45:00,stw,16.77, 12018-10-23 16:01:00,stw,14.74, 0 Periodic Review Meeting
Phase 1: processing at Edge CEP at Edge At Edge CEP1 2018-10-23 08:00:00,stw,15.062018-10-23 06:36:00,stw,15.162018-10-23 15:56:00,stw,14.92018-10-23 15:57:00,stw,14.922018-10-24 08:30:00,stw,16.182018-12-21 18:42:00,stw,17.112018-12-21 18:43:00,stw,16.922018-12-21 18:45:00,stw,16.772018-10-23 16:01:00,stw,14.74 Sensor data Sensor data 2018-10-23 08:00:00,stw,15.06, 02018-10-23 06:36:00,stw,15.16, 02018-10-23 15:56:00,stw,14.9, 02018-10-23 15:57:00,stw,14.92, 12018-10-24 08:30:00,stw,16.18. 02018-12-21 18:42:00,stw,17.11, 02018-12-21 18:43:00,stw,16.92, 02018-12-21 18:45:00,stw,16.77, 12018-10-23 16:01:00,stw,14.74, 0 Added columm: Alarm Info Periodic Review Meeting
Phase 1: processing at Edge CEP at Edge At Edge CEP1 2018-10-23 08:00:00,stw,15.062018-10-23 06:36:00,stw,15.162018-10-23 15:56:00,stw,14.92018-10-23 15:57:00,stw,14.922018-10-24 08:30:00,stw,16.182018-12-21 18:42:00,stw,17.112018-12-21 18:43:00,stw,16.922018-12-21 18:45:00,stw,16.772018-10-23 16:01:00,stw,14.74 Sensor data Sensor data 2018-10-23 08:00:00,stw,15.06, 02018-10-23 06:36:00,stw,15.16, 02018-10-23 15:56:00,stw,14.9, 02018-10-23 15:57:00,stw,14.92, 12018-10-24 08:30:00,stw,16.18. 02018-12-21 18:42:00,stw,17.11, 02018-12-21 18:43:00,stw,16.92, 02018-12-21 18:45:00,stw,16.77, 12018-10-23 16:01:00,stw,14.74, 0 Added columm: Alarm Info Periodic Review Meeting
Phase 2: data ingestion in Data Center Data streams Testbed Gateway Periodic Review Meeting
Phase 2: data ingestion in Data Center 2018-10-23 08:00:00,stw,15.06, 02018-10-23 06:36:00,stw,15.16, 02018-10-23 15:56:00,stw,14.9, 02018-10-23 15:57:00,stw,14.92, 12018-10-24 08:30:00,stw,16.18. 02018-12-21 18:42:00,stw,17.11, 02018-12-21 18:43:00,stw,16.92, 02018-12-21 18:45:00,stw,16.77, 12018-10-23 16:01:00,stw,14.74, 0 Data streams Sensor data Testbed Gateway Periodic Review Meeting
Phase 2: data ingestion in Data Center CEP at Data Center 2018-10-23 08:00:00,stw,15.06, 02018-10-23 06:36:00,stw,15.16, 02018-10-23 15:56:00,stw,14.9, 02018-10-23 15:57:00,stw,14.92, 12018-10-24 08:30:00,stw,16.18. 02018-12-21 18:42:00,stw,17.11, 02018-12-21 18:43:00,stw,16.92, 02018-12-21 18:45:00,stw,16.77, 12018-10-23 16:01:00,stw,14.74, 0 Data streams Sensor data Testbed Gateway Periodic Review Meeting
Phase 2: data ingestion in Data Center CEP at Data Center 2018-10-23 08:00:00,stw,15.06, 02018-10-23 06:36:00,stw,15.16, 02018-10-23 15:56:00,stw,14.9, 02018-10-23 15:57:00,stw,14.92, 12018-10-24 08:30:00,stw,16.18. 02018-12-21 18:42:00,stw,17.11, 02018-12-21 18:43:00,stw,16.92, 02018-12-21 18:45:00,stw,16.77, 12018-10-23 16:01:00,stw,14.74, 0 Data streams Sensor data Testbed Gateway Complex rules check Periodic Review Meeting
Phase 2: data ingestion in Data Center CEP at Data Center 2018-10-23 08:00:00,stw,15.06, 02018-10-23 06:36:00,stw,15.16, 02018-10-23 15:56:00,stw,14.9, 02018-10-23 15:57:00,stw,14.92, 12018-10-24 08:30:00,stw,16.18. 02018-12-21 18:42:00,stw,17.11, 02018-12-21 18:43:00,stw,16.92, 02018-12-21 18:45:00,stw,16.77, 12018-10-23 16:01:00,stw,14.74, 0 Data streams Sensor data Testbed Gateway Complex rules check LeanXcale DB Business Agreement Data KiVi storage system Periodic Review Meeting
Phase 2: data ingestion in Data Center CEP at Data Center 2018-10-23 08:00:00,stw,15.06, 02018-10-23 06:36:00,stw,15.16, 02018-10-23 15:56:00,stw,14.9, 02018-10-23 15:57:00,stw,14.92, 12018-10-24 08:30:00,stw,16.18. 02018-12-21 18:42:00,stw,17.11, 02018-12-21 18:43:00,stw,16.92, 02018-12-21 18:45:00,stw,16.77, 12018-10-23 16:01:00,stw,14.74, 0 Data streams Sensor data Sensor data Testbed Gateway 2018-10-23 08:00:00,stw,15.06, 02018-10-23 06:36:00,stw,15.16, 02018-10-23 15:56:00,stw,14.9, 02018-10-23 15:57:00,stw,14.92, 12018-10-24 08:30:00,stw,16.18. 12018-12-21 18:42:00,stw,17.11, 02018-12-21 18:43:00,stw,16.92, 02018-12-21 18:45:00,stw,16.77, 12018-10-23 16:01:00,stw,14.74, 0 Complex rules check LeanXcale DB KiVi storage system Business Agreement Data Periodic Review Meeting
Phase 2: data ingestion in Data Center CEP at Data Center 2018-10-23 08:00:00,stw,15.06, 02018-10-23 06:36:00,stw,15.16, 02018-10-23 15:56:00,stw,14.9, 02018-10-23 15:57:00,stw,14.92, 12018-10-24 08:30:00,stw,16.18. 02018-12-21 18:42:00,stw,17.11, 02018-12-21 18:43:00,stw,16.92, 02018-12-21 18:45:00,stw,16.77, 12018-10-23 16:01:00,stw,14.74, 0 Data streams Sensor data Sensor data Sensor data Testbed Gateway 2018-10-23 08:00:00,stw,15.06, 02018-10-23 06:36:00,stw,15.16, 02018-10-23 15:56:00,stw,14.9, 02018-10-23 15:57:00,stw,14.92, 12018-10-24 08:30:00,stw,16.18. 12018-12-21 18:42:00,stw,17.11, 02018-12-21 18:43:00,stw,16.92, 02018-12-21 18:45:00,stw,16.77, 12018-10-23 16:01:00,stw,14.74, 0 Complex rules check LeanXcale DB Business Agreement Data KiVi storage system Anomaly table Periodic Review Meeting
Phase 3: Data Quality Assessment Added Data qualitycolumn Data Quality Assessment Sensor data 2018-10-23 08:00:00,stw,15.06, 0, 0.992018-10-23 06:36:00,stw,15.16, 0, 0.972018-10-23 15:56:00,stw,14.9, 0, 0.942018-10-23 15:57:00,stw,14.92, 1, 0.332018-10-24 08:30:00,stw,16.18. 1, 0.422018-12-21 18:42:00,stw,17.11, 0, 0.722018-12-21 18:43:00,stw,16.92, 0, 0.992018-12-21 18:45:00,stw,16.77, 1, 0.662018-10-23 16:01:00,stw,14.74, 0. 0.98 . 2018-10-23 08:00:00,stw,15.06, 02018-10-23 06:36:00,stw,15.16, 02018-10-23 15:56:00,stw,14.9, 02018-10-23 15:57:00,stw,14.92, 12018-10-24 08:30:00,stw,16.18. 12018-12-21 18:42:00,stw,17.11, 02018-12-21 18:43:00,stw,16.92, 02018-12-21 18:45:00,stw,16.77, 12018-10-23 16:01:00,stw,14.74, 0 KiVi storage system LeanXcale DB Periodic Review Meeting
Phase 4: Preventive Maintenance Preventive Maintenance ML. 2018-10-23 08:00:00,stw,15.06, 0, 0.992018-10-23 06:36:00,stw,15.16, 0, 0.972018-10-23 15:56:00,stw,14.9, 0, 0.942018-10-23 15:57:00,stw,14.92, 1, 0.332018-10-24 08:30:00,stw,16.18. 1, 0.422018-12-21 18:42:00,stw,17.11, 0, 0.722018-12-21 18:43:00,stw,16.92, 0, 0.992018-12-21 18:45:00,stw,16.77, 1, 0.662018-10-23 16:01:00,stw,14.74, 0. 0.98 . Detected anomaly KiVi storage system LeanXcale DB Anomaly table: Ano1, …. Ano2, …. Periodic Review Meeting
Phase 5: Danaos Alert Detection Daemon Danaos daemon KiVi storage system LeanXcale DB Anomaly table: Ano1, …. Ano2, …. Periodic Review Meeting
Seamless Query path Seamless Component Periodic Review Meeting
Seamless Query path SQL Query: SELECT id from emp WHERE emp = ‘007’ LXSFederator Seamless Component Periodic Review Meeting
Seamless Query path SQL Query: SELECT id from emp WHERE emp = ‘007’ LXSFederator LXSDB Seamless Component Periodic Review Meeting
Seamless Query path SQL Query: SELECT id from emp WHERE emp = ‘007’ LXSFederator Spark JDBC Thrift Connector LXSDB with Data Skipping Object Storage Seamless Component Periodic Review Meeting