Using the low-latency Apache Pulsar we can build up millions of streams of concurrent data and join them in real time with Apache Flink. We need an ultra-low latency database that can support these workloads to build next-generation IoT, financial and instant analytical transit applications By sinking data into ScyllaDB we enable amazingly fast applications that can grow to any size and join with existing data sources. The next generation of apps is being built now, you must choose the right low-latency scalable platform for these massively data-intensive applications. ScyllaDB + Pulsar + Flink is that platform. Choose Open, Choose Fast, and Make the right choice.

1. Sink Your Teeth
into Streaming
at Any Scale
David Kjerrumgaard, Developer Advocate at Stream Native
Tim Spann, Developer Advocate at Stream Native

2. David Kjerrumgaard
■ Apache Pulsar Committer | Author of Pulsar In Action
■ Former Principal Software Engineer on Splunk’s
messaging team that is responsible for Splunk’s internal
Pulsar-as-a-Service platform.
■ Former Director of Solution Architecture at Streamlio.
Your photo
goes here,
smile :)

3. Tim Spann
■ FLiP(N) Stack = Flink, Pulsar and NiFi Stack
■ Streaming Systems & Data Architecture Expert
■ DZone Iot and ML Top Expert
■ 15+ years of experience with streaming technologies including Pulsar,
Flink, Spark, NiFi, Big Data, Cloud, MXNet, IoT, Python and more.
■ Today, he helps to grow the Pulsar community sharing rich technical
knowledge and experience at both global conferences and through
individual conversations.

4. ■ The Team to Stream
■ What is Apache Pulsar?
■ Join The Streams with Flink SQL
■ Fast Sink to ScyllaDB
■ Reference Architecture
■ Demo
Presentation Agenda

5. Stream Team

6. Apps
Building Real-Time Requires a Team

7. Apache Pulsar

8. Apache Pulsar is a Cloud-Native Messaging and
Event-Streaming Platform.

9. CREATED
Originally developed
inside Yahoo! as
Cloud Messaging
Service
GROWTH
10x Contributors
10MM+ Downloads
Ecosystem Expands
Kafka on Pulsar
AMQ on Pulsar
Functions
. . .
2012 2016 2018 TODAY
APACHE TLP
Pulsar becomes
Apache top level
project.
OPEN SOURCE
Pulsar committed
to open source.
Apache Pulsar Timeline

10. Pulsar Growth
Contributors
11X Growth
Github Stars
5X Growth
10,000,000+
Docker Images
Downloads

11. Streaming
Consumer
Consumer
Consumer
Subscription
Shared
Failover
Consumer
Consumer
Subscription
In case of failure in
Consumer B-0
Consumer
Consumer
Subscription
Exclusive
X
Consumer
Consumer
Key-Shared
Subscription
Pulsar
Topic/Partition
Messaging

12. Multi-Tenancy Model
Tenants
(Data Services)
Namespace
(Microservices)
Pulsar Cluster
Tenants
(Marketing)
Tenants
(Compliance)
Namespace
(ETL)
Namespace
(Campaigns)
Namespace
(ETL)
Namespace
(Risk Assessment)
Topic-1
(Cust Auth)
Topic-1
(Location Resolution)
Topic-2
(Demographics)
Topic-1
(Budgeted Spend)
Topic-1
(Acct History)
Topic-2
(Risk Detection)

13. Integrated Schema Registry
Schema Registry
schema-1 (value=Avro/Protobuf/JSON) schema-2
(value=Avro/Protobuf/JSON)
schema-3
(value=Avro/Protobuf/JSON)
Schema
Data
ID
Local Cache
for Schemas
+
Schema
Data
ID +
Local Cache
for Schemas
Send schema-1
(value=Avro/Protobuf/JSON) data
serialized per schema ID
Send (register)
schema (if not in
local cache)
Read schema-1
(value=Avro/Protobuf/JSON) data
deserialized per schema ID
Get schema by ID (if
not in local cache)
Producers Consumers

14. Automatic Load Balancing
Pulsar supports automatic load shedding whenever the system recognizes a
particular broker is overloaded, the system forces some traffic to be reassigned
to less-loaded brokers.

15. Key Features Apache Pulsar Apache Kafka
Message retention (time-based) ✔ ✔
Message replay ✔ ✔
Message retention (acknowledge-based) ✔ ✘
Built-in tiered storage ✔ ✘
Processing capabilities (fully managed) Pulsar Functions KStream
Queue semantics (round robin) ✔ ✘
Queue semantics (key based) ✔ ✘
Dead letter queue ✔ ✘

16. Key Features Apache Pulsar Apache Kafka
Scheduled and delayed delivery ✔ ✔
Rebalance-free scaling ✔ ✔
Elastically scalable ✔ ✘
Maximum number of topics Millions Up to 100k
Built-in multi-tenancy ✔ ✘
Built-in geo-replication ✔ ✘
Built-in schema management ✔ ✘
End-to-end encryption ✔ ✘

17. Flink SQL

18. Pulsar-Flink Connector
■ Pulsar provides a connector that enables Flink to read & write data directly from/to Pulsar
topics.
■ This makes Pulsar Topic data available for Flink SQL queries.
■ https://hub.streamnative.io/data-processing/pulsar-flink/1.15.1.1/

19. Flink SQL

20. Apache Flink
■ Unified computing engine
■ Batch processing is a special case
of stream processing
■ Stateful processing
■ Massive Scalability
■ Flink SQL for queries, inserts
against Pulsar Topics
■ Streaming Analytics
■ Continuous SQL
■ Continuous ETL
■ Complex Event Processing
■ Standard SQL Powered by Apache
Calcite
Apache Flink

21. Streaming Tables
■ A streaming table represents an unbounded sequence of structured data, i.e.
“facts”
■ Facts in a streaming table are immutable, which means new events can be
inserted into a table, but existing events can never be updated or deleted.
■ All the topics within a Pulsar namespace will automatically be mapped to
streaming tables in a catalog configured to use a pulsar connector.
■ Streaming tables can also be created or deleted via DDL queries, where the
underlying Pulsar topics will be created or deleted.

22. Sink to ScyllaDB

23. Data Offloaders
(Tiered Storage)
Client Libraries
StreamNative Pulsar ecosystem
hub.streamnative.io
Connectors
(Sources & Sinks)
Protocol Handlers
Pulsar Functions
(Lightweight Stream
Processing)
Processing Engines
… and more!
… and more!

24. ScyllaDB Compatible Sink Connector
pulsar.apache.org/docs/en/io-quickstart/
ScyllaDB Compatible Sink

25. David’s ScyllaDB Sink Update
Connector automatically interrogates the database to
determine the schema type definition at runtime.
Includes a framework that extracts the values from the generic
schema types (GenericRecord, and JSON String)
Performance improvements & Bug Fixes
PR: https://github.com/apache/pulsar/pull/16179

26. Building Real-Time Together
Easily integrate ScyllaDB’s fast distributed database system and the StreamNative streaming
platform with Pulsar sources and sinks via various protocols and libraries.
Create high-speed streaming pipelines for log analysis, IoT sensing, and more.

27. Reference Architecture

28. Reference Architecture
Microservices
ETL

29. Reference Architecture
ScyllaDB <-> Apache Pulsar -> Flink SQL -> Apache Pulsar -> Apache Pinot
Flink SQL for Continuous Analytics, Ingest, Real-time Joins, Real-Time Analytics
Apache Pulsar for routing, transformation, central messaging hub, data channels
ScyllaDB for instant applications and massive data feeds
Apache Pinot for low latency instant result queries

30. Thank You
Stay in Touch
Timothy Spann
tim@streamnative.io
@PaaSDev
https://github.com/tspannhw
https://www.linkedin.com/in/timothyspann/
David Kjerrumgaard
david@streamnative.io
@DavidKjerrumga1
https://github.com/david-streamlio
https://www.linkedin.com/in/davidkj/