A presentation at Conf42: Machine Learning 2021 by Tim Spann
Hail Hydrate! From Stream to Lake Timothy Spann Developer Advocate https://github.com/tspannhw/SpeakerProfile
https://github.com/tspannhw https://www.datainmotion.dev/
Speaker Bio Developer Advocate DZone Zone Leader and Big Data MVB; @PaasDev https://github.com/tspannhw https://www.datainmotion.dev/ https://github.com/tspannhw/SpeakerProfile https://dev.to/tspannhw https://sessionize.com/tspann/ https://www.slideshare.net/bunkertor
AGENDA Use Case - Populate the Data Lake Key Challenges ▪ Their Impact ▪ A Solution ▪ Outcome Why Apache NiFi and Apache Pulsar? Successful Architecture Demo Next Steps
USE CASE IoT Ingestion: High-volume streaming sources, multiple message formats, diverse protocols and multi-vendor devices creates data ingestion challenges. 5
KEY CHALLENGES Data Ingestion: High-volume streaming sources, multiple message formats, diverse protocols and multi-vendor devices creates data ingestion challenges. Real-time Insights: Analyzing continuous and rapid inflow (velocity) of streaming data at high volumes creates major challenges for gaining real-time insights. Visibility: Lack visibility of end-to-end streaming data flows, inability to troubleshoot bottlenecks, consumption patterns etc. 6
IMPACT Code Sprawl: Custom scripts over various qualities proliferate across environments to cope with the complexity. Costs: Increasing costs of development and maintenance. Too many tools, not enough experts, waiting for contractors or time delays as developers learn yet another tool, package or language. Delays: Decreasing user satisfaction and delay in project delivery. Missed revenue and opportunities. 7
SOLUTION Data Ingestion: Apache NiFi is the one tool handle high-volume streaming sources, multiple message formats, diverse protocols and multi-vendor devices. Variety of Data: Apache NiFi offers hundreds of OOTB connectors and a GUI that accelerates flow developments. With Record Processors that convert types in a single fast step. Visibility: Apache NiFi provenance provides insights, metrics and control over the entire end-to-end stream across clouds. 8
OUTCOME New Applications: Enablement of new innovative use cases in compressed timeframe. No more waiting for data to arrive, Data Analysts and Data Scientists focus on innovation. Savings: Cost reduction thanks to technologies offload, reduced consultant costs and simplification of ingest processes. Agility: Reduction of new data source onboarding time from weeks to days. More data in your data warehouse now. 9
FLiP Stack for Cloud Data Engineers - ML Multiple users, frameworks, languages, clouds, data sources & clusters CLOUD DATA ENGINEER • Experience in ETL/ELT CAT AI / Deep Learning / ML / DS • Expert in ETL (Eating, Ties and Laziness) • Edge Camera Interaction • Typical User • Can run in Apache NiFi • No Coding Skills • Can run in Apache Flink • Experience with Streaming • Can use NiFi • Knowledge of Cloud Tools • Questions your cloud spend • Can run in Apache NiFi - MiNiFi Agents • Coding skills in Python or Java • Knowledge of database query languages such as SQL • Can run in Apache Pulsar Functions
FLiP Stack (FLink -integrate- Pulsar) https://hub.streamnative.io/data-processing/pulsar-flink/2.7.0/
WHAT IS APACHE NIFI? Apache NiFi is a scalable, real-time streaming data platform that collects, curates, and analyzes data so customers gain key insights for immediate actionable intelligence. 12
APACHE NIFI Enable easy ingestion, routing, management and delivery of any data anywhere (Edge, cloud, data center) to any downstream system with built in end-to-end security and provenance ACQUIRE • Advanced tooling to industrialize flow development (Flow Development Life Cycle) • • • • DELIVER PROCESS FTP HASH ENCRYPT GEOENRICH FTP SFTP MERGE TALL SCAN SFTP HL7 EXTRACT EVALUATE REPLACE HL7 UDP DUPLICATE EXECUTE TRANSLATE UDP XML SPLIT CONVERT XML HTTP ROUTE TEXT HTTP EMAIL ROUTE CONTENT EMAIL HTML ROUTE CONTEXT HTML IMAGE CONTROL RATE IMAGE SYSLOG DISTRIBUTE LOAD SYSLOG Over 300 Prebuilt Processors Easy to build your own Parse, Enrich & Apply Schema Filter, Split, Merger & Route Throttle & Backpressure • • • • Guaranteed Delivery Full data provenance from acquisition to delivery Diverse, Non-Traditional Sources Eco-system integration
WHAT IS APACHE PULSAR? Apache Pulsar is an open source, cloud-native distributed messaging and streaming platform. EVENTS 14
APACHE PULSAR Enable Geo-Replicated Messaging ● ● ● ● ● ● ● ● ● ● ● ● Pub-Sub Geo-Replication Pulsar Functions Horizontal Scalability Multi-tenancy Tiered Persistent Storage Pulsar Connectors REST API CLI Many clients available Four Different Subscription Types Multi-Protocol Support ○ MQTT ○ AMQP ○ JMS ○ Kafka ○ …
APACHE FLINK 3B+ data points daily streaming in from 25 million customers running real time machine learning prediction USE CASE Streaming real-time data pipelines that need to handle complex stream or batch data event processing, analytics, and/or support event-driven applications TECHNOLOGY Flink performs compute at in-memory speed at any scale Flink parses SQL using Apache Calcite, which supports standard ANSI SQL Flink runs standalone, on YARN, and has a K8s Operator APPLICATION Comcast a global media uses Flink for operationalizing machine learning models and near-real-time event stream processing Flink helps deliver a personalized, contextual interaction reducing time to support resolutions saving millions of dollars per year Flink CONSIDERATION Data Freshness SLAs Flink can read and write from Hive data Review requirements for fault tolerance, resilience, and HA
Apache MXNet Native Processor through DJL.AI for Apache NiFi This processor uses the DJL.AI Java Interface https://github.com/tspannhw/nifi-djl-processor https://dev.to/tspannhw/easy-deep-learning-in-apache-nifi-with-djl-2d79
Apache MXNet Native Processor for Apache NiFi ● https://www.slideshare.net/bunkertor/apache-deep-learning-101-apachecon-montreal-2018-v031 ● https://www.slideshare.net/bunkertor/apache-deep-learning-202-washington-dc-dws-2019 ● https://www.slideshare.net/bunkertor/apache-deep-learning-201-barcelona-dws-march-2019
Apache OpenNLP with Apache NiFi Apache OpenNLP for Entity Resolution Processor https://github.com/tspannhw/nifi-nlp-processor Requires installation of NAR and Apache OpenNLP Models (http://opennlp.sourceforge.net/models-1.5/). This is a non-supported processor that I wrote and put into the community. You can write one too! https://community.cloudera.com/t5/Community-Articles/Open-NLP-Example-Apache-NiFi-Processor/ta-p/249293 https://opennlp.apache.org/news/release-190.html
Multiinges t Multiinges t ALL DATA - ANYTIME - ANYWHERE - ANY CLOUD Multi-i ngest Priority Merge
SHOW ME SOME DATA {“uuid”: “rpi4_uuid_jfx_20200826203733”, “amplitude100”: 1.2, “amplitude500”: 0.6, “amplitude1000”: 0.3, “lownoise”: 0.6, “midnoise”: 0.2, “highnoise”: 0.2, “amps”: 0.3, “ipaddress”: “192.168.1.76”, “host”: “rp4”, “host_name”: “rp4”, “macaddress”: “6e:37:12:08:63:e1”, “systemtime”: “08/26/2020 16:37:34”, “endtime”: “1598474254.75”, “runtime”: “28179.03”, “starttime”: “08/26/2020 08:47:54”, “cpu”: 48.3, “cpu_temp”: “72.0”, “diskusage”: “40219.3 MB”, “memory”: 24.3, “id”: “20200826203733_28ce9520-6832-4f80-b17d-f36c21fd8fc9”, “temperature”: “47.2”, “adjtemp”: “35.8”, “adjtempf”: “76.4”, “temperaturef”: “97.0”, “pressure”: 1010.0, “humidity”: 8.3, “lux”: 67.4, “proximity”: 0, “oxidising”: 77.9, “reducing”: 184.6, “nh3”: 144.7, “gasKO”: “Oxidising: 77913.04 Ohms\nReducing: 184625.00 Ohms\nNH3: 144651.47 Ohms”} 21
Weather Streaming Pipeline
DEEPER CONTENT ● ● ● ● ● ● ● https://www.datainmotion.dev/2020/10/running-flink-sql-against-kafka-using.html https://www.datainmotion.dev/2020/10/top-25-use-cases-of-cloudera-flow.html https://github.com/tspannhw/EverythingApacheNiFi https://github.com/tspannhw/CloudDemo2021 https://github.com/tspannhw/StreamingSQLExamples https://www.linkedin.com/pulse/2021-schedule-tim-spann/ https://github.com/tspannhw/StreamingSQLExamples/blob/8d02e62260e82b027b43abb911b5c366 a3081927/README.md 23
TH NK YOU
29-july-2021 - Conf42 - Machine Learning - Hail Hydrate! From Stream to Lake Apache NiFi, Apache Flink, Apache Pulsar https://www.conf42.com/Machine_Learning_2021_Tim_Spann_From_Stream_to_Lake
https://github.com/tspannhw/SpeakerProfile/tree/main/2021/talks
https://www.youtube.com/watch?v=CfCp_xPvxCg&list=PLIuxSyKxlQrD02X7IKNNxFMy_K8oEejAu&index=22
for free. You
can too.