DataLake POC



As DataLake infrastructure itself will be considered outside of ONAP scope, the POC will focus on delivering one of key component of DL-Feeder as service under DCAE project.



Architecture







POC Scope

  • DataLake Feeder and DLAdmin delivery as services under DCAE  (dcaegen2/services repository)
     

    • dcaegen2/services

      └── components

          ├── dl-handler

          │   ├── dl-admin

          │   └── dl-feeder

  • DLFeeder will support pluggable sources and datastore support

    • DMAAP/Kafka as primary source (option to poll kafka broker for new topic?)

    • Druid/Couchbase/Elasticsearch/Hadoop datastore support

  • ConfigBingingService integration to fetch deployment/run-time configuration

  •  

    • Alternate option will be to leverage k8s Configmap.  

  • Integrate the DL Admin UI into DCAE dashboard, which is built on portal SDK and serves as UAM for DCAE users/operation.

  • Integrate DLAdmin into DCAE-Dashboard



Other extension cases (stretch goal)

  • RDMS datastore support

  • Support other input data stream beside Dmaap/kafka

  • Integrate with PNDA/datastore

  • DCAE Platform integration for dynamic topic configuration

  • Logging project integration



Deliverable

  • DL Feeder container

  • DL Admin GUI container

  • Documentation (wiki and readme files) to setup/configure above

  • Instruction for Test storage and UI setup and DL configuration management to demo data flow and visualization



Resources (To be updated)

China Mobile:

  • Guobiao Mo

  • Tao Shen

  • Binbin Tian

  • Man Liu

QCT

  • Karl Chiang

  • Kate Hsuan

  • Ekko Chang

  • May Lin

IBM

  • Driptaroop Das