EXPAND FOR MORE INFO.
Speaker: Gokul Gunasekaran, Cask
Big Data Applications Meetup, 06/15/2016
Palo Alto, CA
More info here: http://www.meetup.com/BigDataApps/
Link to video:
About the talk:
Cask Hydrator is an extension to the open source Cask Data Application Platform (CDAP) that simplifies the process of developing and operating realtime and batch data pipelines on Hadoop. Hydrator’s web-based drag-and-drop UI allows users to quickly build hadoop-scalable, distro-agnostic data pipelines without writing any code.
Powered by CDAP (http://cdap.io), Hydrator provides ease of operability through metadata information, lineage, metrics and log collection in a single location. In this talk, we will build data pipelines, with real-life applications, that pull in data from multiple sources, train and use a machine learning model to classify data using Spark MLLib, and write data to different sinks. We will also delve under the covers to see how these data pipelines are transformed to a series of MapReduce/Spark jobs and also touch upon some interesting challenges we had to tackle while developing Hydrator.
DEV meet-up UiPath Document Understanding May 7 2024 Amsterdam
Building Data pipelines with Cask Hydrator, by Gokul Gunasekaran from Cask
1. Building Data Pipelines with Cask Hydrator
Gokul Gunasekaran
Software Engineer, Cask Data
June 15, 2016
Cask, CDAP, Cask Hydrator and Cask Tracker are trademarks or registered trademarks of Cask Data. Apache Spark, Spark, the Spark logo, Apache Hadoop, Hadoop and the Hadoop logo are trademarks or registered trademarks of the Apache Software Foundation. All other trademarks and registered trademarks are the property of their respective owners.
2. cask.co
INGEST
any data from any source
in real-time and batch
BUILD
drag-and-drop ETL/ELT
pipelines that run on Hadoop
EGRESS
any data to any destination
in real-time and batch
Data Pipeline
provides the ability to automate complex workflows that involves fetching data,
performing non-trivial transformations, deriving and serving insights from the data
2
3. cask.co
Web Analytics and Reporting Use Case
✦ Hadoop ETL pipeline(s) stitched together using hard-to-maintain, brittle scripts
✦ Not many developers with expertise in Hadoop components (HDFS, MapReduce, Spark, YARN,
HBase, Kafka)
✦ Hard to debug and validate, resulting in frequent failures in production environment
Fetch web access logs from S3 every hour, load it into Hadoop cluster for backup and perform
analytics and enable realtime reporting of no. of successful/failure responses and client browser info
Challenge —
3
4. cask.co
Demo
Load Log Files from S3 into
HDFS and perform
aggregations/analysis
• Start with web access logs stored in Amazon S3
• Store the raw logs into HDFS Avro Files
• Parse the access log lines into individual fields
• Find out distribution of status codes
• Find out the most commonly used client browser
4
6. cask.co
Hydrator Studio
✦ Drag-and-drop GUI for visual Data
Pipeline creation
✦ Rich library of pre-built sources,
transforms, sinks for data ingestion and
ETL use cases
✦ Separation of pipeline creation from
execution framework - MapReduce,
Spark, Spark Streaming etc.
✦ Hadoop-native and Hadoop Distro
agnostic
6
7. cask.co
Hydrator Data Pipeline
✦ Captures Metadata, Audit, Lineage
info and visualized using Cask
Tracker
✦ Post-run notification, centralized
metrics and log collection for ease of
operability
✦ Simple Java API to build your own
source, transforms, sinks with class
loader isolation
✦ SparkML based plugins, Python
transforms for data scientists
7
8. cask.co
✦ ElasticSearch, Cassandra, Kafka, SFTP, JMS and many more sources and sinks
✦ De-duplicate, Group By Aggregation, Row Denormalizer and other transforms
Out of the box Integrations
8
9. cask.co
✦ Implement your own batch (or realtime) source, transform, sink plugins using simple Java API
Custom Plugins
9
11. cask.co
Pipeline Implementation
Logical Pipeline
Physical Workflow
MR/Spark Executions
Planner
CDAP
✦ Planner converts logical pipeline to a physical
execution plan
✦ Optimizes and bundles functions into one or more
MR/Spark jobs
✦ CDAP is the runtime environment where all the
components of the data pipeline are executed
✦ CDAP provides centralized log and metrics collection,
transaction, lineage and audit information
11
12. cask.co
✦ Join across multiple data sources (CDAP-5588)
✦ Pipeline preview
✦ Macro substitutions
✦ Pre-Actions in pipelines similar to post run notifications
✦ Spark streaming support for Realtime pipelines
Upcoming capabilities
12
14. cask.co
Self-Service Data Ingestion
and ETL for Data Lakes
Built for Production
on CDAP
Rich Drag-and-Drop
User Interface
Open Source &
Highly Extensible
14