Data pipeline architecture for cloud processing of structured and unstructured data

• US 10,235,076 B2
• Filed: 12/14/2016
• Issued: 03/19/2019
• Est. Priority Date: 12/16/2015
• Status: Active Grant

• Current Assignee: Accenture Global Solutions Limited (Accenture PLC)
• Original Assignee: Accenture Global Solutions Limited (Accenture PLC)
• Inventors: Das, Jagaran, Tung, Teresa Sheausan, Yelisetty, Srinivas, Corin, Daniel
• Primary Examiner(s): Woolwine, Shane D
• Application Number:
  US15/378,244
• Publication Number:
  US 20170177263A1
• Time in Patent Office: 825 Days
• Field of Search: 711117, 711157, 711158
• CPC Class Codes: G06F 13/382 using universal interface adapter
• CPC Class Codes: G06F 13/4068 Electrical coupling
• CPC Class Codes: G06F 16/24568 Data stream processing; Continuous queries
• CPC Class Codes: G06F 3/0604 Improving or facilitating administration, e.g. storage management
• CPC Class Codes: G06F 3/0613 in relation to throughput
• CPC Class Codes: G06F 3/0635 by changing the path, e.g. traffic rerouting, path reconfiguration
• CPC Class Codes: G06F 3/0638 Organizing or formatting or addressing of data
• CPC Class Codes: G06F 3/0683 Plurality of storage devices
• CPC Class Codes: G06F 3/0685 Hybrid storage combining heterogeneous device types, e.g. hierarchical storage, hybrid arrays
• CPC Class Codes: G06F 9/5072 Grid computing

Abstract

Scalable architectures provide resiliency and redundancy and are suitable for cloud deployment. The architectures support extreme data throughput requirements. In one implementation, the architectures provide a serving layer and an extremely high speed processing lane. With these and other features, the architectures support complex analytics, visualization, rule engines, and centralized pipeline configuration.

Claims

1. A system comprising:

a data stream processing pipeline comprising sequential multiple processing stages, including;

an integration stage comprising a connector interface configurable to connect multiple incoming data streams from multiple different data stream input sources to multiple different ingestion processors in an ingestion stage; and

a storage stage comprising a memory hierarchy with multiple different memory bandwidths, the memory hierarchy comprising;

a database repository configured to support a batch target processing bandwidth of the multiple different memory bandwidths;

an in-memory data store configured to support an in-memory target processing bandwidth of the multiple different memory bandwidths, the in-memory target processing bandwidth being faster than the batch target processing bandwidth; and

pipeline configuration circuitry operable to;

read a pipeline configuration for data stream processing through the data stream processing pipeline; and

configure the multiple processing stages according to the pipeline configuration to accept and process an individual incoming data stream of the multiple incoming data streams for delivery to the memory hierarchy according to a selected target processing bandwidth of the multiple different memory bandwidths, the selected target processing bandwidth defined for the individual incoming data stream.

2. The system of claim 1, where:

the pipeline configuration is configured to direct a processed data stream from the ingestion stage into the in-memory data store when the target processing bandwidth exceeds a bandwidth threshold.

3. The system of claim 2, where:

the pipeline configuration is configured to direct the processed data stream into the database repository for a batch target processing bandwidth.

4. The system of claim 2, where:

the pipeline configuration is configured to direct the processed data stream into the in-memory data store when the selected target processing bandwidth is faster than batch processing.

5. The system of claim 1, where:

the multiple processing stages include a data stream processing stage; and

the pipeline configuration specifies a particular data stream processor available within the data stream processing stage.

6. The system of claim 5, where:

the multiple processing stages include a visualization stage; and

the pipeline configuration specifies a particular visualization processor available within the visualization stage.

7. The system of claim 1, further comprising:

a reactive interface defined between the ingestion stage and the storage stage that establishes a message subscription relationship between the storage stage and the ingestion stage.

8. The system of claim 1, where:

the pipeline configuration comprises a pipeline configuration file that comprises references to external pipeline stage configuration files separate from the pipeline configuration file itself.

9. A method comprising:

in a data processing architecture;

reading a pipeline configuration for data stream processing through a multiple stage data stream processing pipeline including an integration stage and an ingestion stage;

configuring, responsive to the pipeline configuration, the integration stage to accept and process multiple incoming data streams from multiple endpoints;

at the ingestion stage, ingesting the multiple incoming streams to create multiple ingested data streams;

routing, responsive to the pipeline configuration, selected ingested data stream of the multiple ingested data streams from the ingestion stage to a memory hierarchy defining data storage options for multiple different target processing bandwidths, the memory hierarchy comprising;

a database repository configured to support a batch target processing bandwidth of the multiple different target processing bandwidths;

an in-memory data store configured to support an in-memory target processing bandwidth of the multiple different target processing bandwidths, the in-memory target processing bandwidth being faster than the batch target processing bandwidth; and

selecting from among the data storage options responsive to a selected target processing bandwidth defined for the selected ingested data stream.

10. The method of claim 9, where:

configuring further comprises;

selecting a particular visualization processor available within a visualization stage in the multiple stage data stream processing pipeline.

11. The method of claim 10, where:

configuring further comprises;

selecting a particular data stream processor available within a data stream processing stage in the multiple stage data stream processing pipeline.

12. The method of claim 11, where:

configuring further comprises;

selecting a particular connector to delivery output of the data stream processor to the visualization processor.

13. The method of claim 9, where:

routing comprising;

routing the ingested data stream to the in-memory data store when a target processing bandwidth for the ingested data stream exceeds a bandwidth threshold.

14. The method of claim 9, where:

routing comprising;

routing the ingested data stream to the database repository when a target processing bandwidth for the ingested data stream is below a bandwidth threshold.

15. A system comprising:

a data stream processing pipeline comprising sequential multiple processing stages, including;

an integration stage comprising a connector interface configurable from multiple different data stream input sources to multiple different ingestion processors;

an ingestion stage comprising the multiple different ingestion processors;

a storage stage comprising a memory hierarchy defining data stores with multiple different memory bandwidths, the memory hierarchy comprising;

a database repository configured to support a batch target processing bandwidth of the multiple different memory bandwidths;

an in-memory data store configured to support an in-memory target processing bandwidth of the multiple different memory bandwidths, the in-memory target processing bandwidth being faster than the batch target processing bandwidth;

a data stream processing stage comprising multiple different data processors;

a visualization stage comprising multiple different visualization processors; and

a service stage operable to connect the data stream processing stage to the visualization stage; and

pipeline configuration circuitry operable to;

read a pipeline configuration file for configuring data stream processing through the data stream processing pipeline; and

configure the multiple processing stages to accept and process an incoming data stream for delivery to the memory hierarchy according to a selected target processing bandwidth defined for the incoming data stream, through the data stream processing stage, and to the visualization stage.

16. The system of claim 15, where:

the pipeline configuration file comprises;

an integration stage section comprising an integration configuration entry for configuring the integration stage by the pipeline configuration circuitry;

an ingestion stage section comprising an ingestion configuration entry for configuring the ingestion stage by the pipeline configuration circuitry;

a storage stage section comprising a storage configuration entry for configuring the storage stage by the pipeline configuration circuitry;

a processing stage section comprising a processing configuration entry for configuring the data stream processing stage by the pipeline configuration circuitry;

a service stage section comprising a service configuration entry for configuring the service stage by the pipeline configuration circuitry; and

a visualization stage section comprising a visualization configuration entry for configuring the visualization stage by the pipeline configuration circuitry.

17. The system of claim 15, where:

the pipeline configuration file further comprises references to external pipeline stage configuration files separate from the pipeline configuration file itself.

18. The system of claim 15, where:

the pipeline configuration is configured to direct a processed data stream from the ingestion stage into the in-memory data store when the selected target processing bandwidth exceeds a bandwidth threshold.

19. The system of claim 15, where:

the pipeline configuration is configured to direct the processed data stream into the database repository for a batch target processing bandwidth.

20. The system of claim 15, where:

the pipeline configuration is configured to direct the processed data stream into the in-memory data store when the selected target processing bandwidth is faster than batch processing.