IoT Ingestion & Analytics using Apache Apex - A Native Hadoop Platform
Download as pptx, pdf1 like886 views
The document discusses IoT data ingestion and analytics using Apache Apex, a Hadoop-native platform that addresses challenges such as data velocity, volume, and variety. It outlines the essential features of Apex, including its ability to perform real-time analytics, fault tolerance, and scalability for processing large data streams through a directed acyclic graph (DAG) model of operators. Additionally, the document provides insights into building applications using Apex, its monitoring capabilities, and resources for further exploration.
IoT Ingestion & Analytics using Apache Apex - A Native Hadoop Platform
- 1. Pramod Immaneni <[email protected]> PPMC Member, Architect @DataTorrent Inc Apr 6th, 2016 A native Hadoop platform IoT Ingestion & Analytics using Apache Apex
- 2. © 2015 DataTorrent IoT (Internet of Things) 2 • Network of objects in physical world with electronic sensors • Collecting and sending data •Consumer • Smart home – Smart bulbs, thermostats, locks • Connected vehicles • Wearables • Industrial • Manufacturing • Retail • Energy
- 3. © 2015 DataTorrent Challenges 3 • Velocity of data • Machines generate data at a much higher pace than humans • They generate data continuously without pause posing processing challenges • Certain aspects of data may become stale if not processed in time • Volume • Large volumes of data for processing and storage • Volume keep increasing because of the exponential growth in devices • Research estimates zettabytes by 2020 • Variety • Types of data – sensor, location, kpi • Formats – csv, json • Delivery mechanisms – files, message queues
- 4. © 2015 DataTorrent What should a solution have 4 • Ingest and perform analytics on data in motion • Handle the incoming throughput while maintaining SLAs • Scale with the volume of data • Fault tolerant so that no data is lost • Flexible to handle data variety and allow custom business logic • Easy to use tools for the user
- 5. © 2015 DataTorrent What is Apex 5 • Platform and runtime engine to enable scalable and fault- tolerant distributed applications • Hadoop native • Process streaming or batch big data • High throughput and low latency • Library of commonly needed business logic • Build any custom business logic in your application
- 6. © 2015 DataTorrent Applications 6 • Distributed processing • Application logic broken into components called operators that run in a distributed fashion across your cluster • Scalable • Operators can be partitioned into multiple instances to handle large volumes • Fault tolerant • Automatically recover from node outages without having to reprocess from beginning • State is preserved • Long running applications • Operators • Use library to build applications quickly • Write your own in Java using the API • Operational insight – DataTorrent RTS • See how each operator is performing and even record data
- 7. © 2015 DataTorrent Apex Platform Overview 7
- 8. © 2015 DataTorrent Apache Malhar Library 8
- 9. © 2015 DataTorrent Application Development Model 9 A Stream is a sequence of data tuples A typical Operator takes one or more input streams, performs computations & emits one or more output streams • Each Operator is YOUR custom business logic in java, or built-in operator from our open source library • Operator has many instances that run in parallel and each instance is single-threaded Directed Acyclic Graph (DAG) is made up of operators and streams Directed Acyclic Graph (DAG) Output Stream Tupl e Tupl e er Operator er Operator er Operator er Operator er Operator er Operator
- 10. © 2015 DataTorrent Native Hadoop Integration 10 • YARN is the resource manager • HDFS used for storing any persistent state
- 11. © 2015 DataTorrent Advanced Windowing Support 11 Application window Sliding window and tumbling window Checkpoint window No artificial latency
- 12. © 2015 DataTorrent Scalability & Partitioning 12 NxM PartitionsUnifier 0 1 2 3 Logical DAG 0 1 2 1 1 Unifier 1 20 Logical Diagram Physical Diagram with operator 1 with 3 partitions 0 Unifier 1a 1b 1c 2a 2b Unifier 3 Physical DAG with (1a, 1b, 1c) and (2a, 2b): No bottleneck Unifier Unifier0 1a 1b 1c 2a 2b Unifier 3 Physical DAG with (1a, 1b, 1c) and (2a, 2b): Bottleneck on intermediate Unifier
- 13. © 2015 DataTorrent Advanced Partitioning 13 0 1a 1b 2 3 4Unifier Physical DAG 0 4 3a2a1a 1b 2b 3b Unifier Physical DAG with Parallel Partition Parallel Partition Container uopr uopr1 uopr2 uopr3 uopr4 uopr1 uopr2 uopr3 uopr4 dopr dopr doprunifier unifier unifier unifier Container Container NICNIC NICNIC NIC Container NIC Logical Plan Execution Plan, for N = 4; M = 1 Execution Plan, for N = 4; M = 1, K = 2 with cascading unifiers Cascading Unifiers 0 1 2 3 4 Logical DAG
- 14. © 2015 DataTorrent Dynamic Partitioning 14 • Partitioning change while application is running ᵒ Change number of partitions at runtime based on stats ᵒ Determine initial number of partitions dynamically • Kafka operators scale according to number of kafka partitions ᵒ Supports re-distribution of state when number of partitions change ᵒ API for custom scaler or partitioner 2b 2c 3 2a 2d 1b 1a1a 2a 1b 2b 3 1a 2b 1b 2c 3b 2a 2d 3a Unifiers not shown
- 15. © 2015 DataTorrent Fault Tolerance 15 • Operator state is checkpointed to a persistent store ᵒ Automatically performed by engine, no additional work needed by operator ᵒ In case of failure operators are restarted from checkpoint state ᵒ Frequency configurable per operator ᵒ Asynchronous and distributed by default ᵒ Default store is HDFS • Automatic detection and recovery of failed operators ᵒ Heartbeat mechanism • Buffering mechanism to ensure replay of data from recovered point so that there is no loss of data • Application master state checkpointed
- 16. © 2015 DataTorrent Processing Guarantees 16 Atleast once • On recovery data will be replayed from a previous checkpoint ᵒ Messages will not be lost ᵒ Default mechanism and is suitable for most applications • Can be used in conjunction with following mechanisms to achieve exactly-once behavior in fault recovery scenarios ᵒ Transactions with meta information, Rewinding output, Feedback from external entity, Idempotent operations Atmost once • On recovery the latest data is made available to operator ᵒ Useful in use cases where some data loss is acceptable and latest data is sufficient Exactly once • At least once + state recovery + operator logic to achieve end-to-end exactly once
- 17. © 2015 DataTorrent Ingestion Operators 17 • NFS/FTP/S3/HDFS Readers and Writers • Kafka readers and writers • Cassandra, HBase, Hive, JDBC • CSV, JSON, Avro, XML parsers • Filtering and enrichment • Dedup • Dimension computation • Stream Joins • …
- 18. © 2015 DataTorrent Pipeline builder 18
- 19. © 2015 DataTorrent Monitoring Console Logical View 19
- 20. © 2015 DataTorrent Monitoring Console Physical View 20
- 21. © 2015 DataTorrent Real-Time Dashboards Real Time Visualization 21
- 22. © 2015 DataTorrent IoT solutions 22
- 23. © 2015 DataTorrent Resources 23 • Ingest application – http://www.datatorrent.com/download • Solutions – http://www.datatorrent.com/solutions • Subscribe - http://apex.incubator.apache.org/community.html • Twitter - @ApacheApex; Follow - https://twitter.com/apacheapex • Free Enterprise License for Startups - https://www.datatorrent.com/product/startup- accelerator/ • Cloud Trial - http://datatorrent.com/cloudtrial.html
- 24. © 2015 DataTorrent We Are Hiring 24 • [email protected] • Developers/Architects • QA Automation Developers • Information Developers • Build and Release • Community Leaders
- 25. © 2015 DataTorrent Upcoming Events 25 • April 14th 6pm PDT – IOT Big Data Ingestion and Processing in Hadoop by Silver Spring Networks • May 4th 6pm PDT – Fault Tolerant HDFS r/w with Apache Apex and Apex (native Hadoop) benchmarks
- 26. End 26