Deploying Driverless AI Models to Production

This documentation lists some of the scenarios to deploy Driverless AI models to production and provides guidlines to create deployment templates for the same.

The final model from a Driverless AI experiment can be exported as either a MOJO scoring pipeline or a Python scoring pipeline. The Mojo scoring pipeline comes with a pipline.mojo file with a Java or C++ runtime. This can be deployed in any environment that supports Java or C++. The Python scoring pipeline comes with the scoring whl file for deployment purposes.

Below, is a list of the deployment scenerios for Driverless AI pipelines for Real-time, Batch or Stream scoring.

graph LR;

  A["Realtime Scoring (single or multi-rows)"]-->AA[DAI MOJO Pipeline with Java Runtime];
  A-->AB[DAI MOJO Pipeline with C++ Runtime];
  A-->AC[DAI PYTHON Scoring Pipeline];

  AA-->AAA["As REST Server"]
  AA-->AAB["As AWS Lambda"]
  AA-->AAC["As GCP Cloud Run"]
  AA-->AAD["As AzureML"]
  AA-->AAE["As library"]
  AA-->AAF["As Apache NiFi "]
  AA-->AAG["As Apache Flink"]

  AB-->ABA["As library"]
  AB-->ABB["As Apache NiFi "]

  AC-->ACA["As REST Server"]
  AC-->ACB["As Apache NiFi"]

  click AAA "./local-rest-scorer"
  click AAB "./aws_lambda_scorer"
  click AAC "./gcp"
  click AAF "https://github.com/james94/dai-deployment-examples/tree/mojo-nifi/mojo-nifi"
  click AAG "https://github.com/h2oai/dai-deployment-examples/tree/master/mojo-flink"

graph LR;

  A["Batch Scoring"]-->AA[DAI MOJO Pipeline with Java Runtime];
  A-->AB[DAI MOJO Pipeline with C++ Runtime];
  A-->AC[DAI PYTHON Scoring Pipeline];

  AA-->AAA["As Apache Spark batch"]
  AA-->AAB["As library"]
  AA-->AAC["As Hive UDF"]
  AA-->AAD["As DB scorer"]
  AA-->AAE["As Apache NiFi"]
  AA-->AAF["As Apache Flink"]
  AA-->AAG["As Snowflake Workflow"]

  AB-->ABA["As library"]
  AB-->ABB["As Apache NiFi"]

  AC-->ACA["As library"]
  AC-->ACB["As Apache NiFi"]

  click AAA "http://docs.h2o.ai/sparkling-water/3.0/latest-stable/doc/deployment/load_mojo_pipeline.html#loading-and-score-the-mojo"
  click AAD "./sql-jdbc-scorer"
  click AAE "https://github.com/james94/dai-deployment-examples/tree/mojo-nifi/mojo-nifi"
  click AAF "https://github.com/h2oai/dai-deployment-examples/tree/master/mojo-flink"
  click AAG "https://www.youtube.com/watch?v=EMpGVer01WE"

graph LR;

  A["Stream Scoring"]-->AA[DAI MOJO Pipeline with Java Runtime];

  AA-->AAA["As Apache Spark Stream"]
  AA-->AAB["As Task Queue "]
  AA-->AAC["As Active MQ"]
  AA-->AAD["As Kafka Topic"]

  click AAD "https://github.com/h2oai/dai-deployment-examples/blob/master/mojo-flink/daimojo-flink-kafka.md"