[DevNexus-2018] Apache Kafka A Streaming Data Platform

Slide 1

Slide 1 text

@ Apache Kafka A Streaming Data Platform

Slide 2

Slide 2 text

@ @gamussa @confluentinc Who am I?

Slide 3

Slide 3 text

@ @gamussa @confluentinc Solutions Architect Who am I?

Slide 4

Slide 4 text

@ @gamussa @confluentinc Solutions Architect Developer Advocate Who am I?

Slide 5

Slide 5 text

@ @gamussa @confluentinc Solutions Architect Developer Advocate @gamussa in internetz Who am I?

Slide 6

Slide 6 text

@ @gamussa @confluentinc Solutions Architect Developer Advocate @gamussa in internetz Hey you, yes, you, go follow me in twitter © Who am I?

Slide 7

Slide 7 text

@ @gamussa @confluentinc

Slide 8

Slide 8 text

@ @gamussa @confluentinc A company is build on

Slide 9

Slide 9 text

@ @gamussa @confluentinc A company is build on DATA FLOWS but All we have is DATA STORES

Slide 10

Slide 10 text

@ @gamussa @confluentinc

Slide 11

Slide 11 text

@ @gamussa @confluentinc

Slide 12

Slide 12 text

@ @gamussa @confluentinc

Slide 13

Slide 13 text

@ @gamussa @confluentinc

Slide 14

Slide 14 text

@ @gamussa @confluentinc

Slide 15

Slide 15 text

@ @gamussa @confluentinc

Slide 16

Slide 16 text

@ @gamussa @confluentinc Streaming Platform 1. Pub/Sub 2. Store 3. Process

Slide 17

Slide 17 text

@ @gamussa @confluentinc Streaming Platform 1. Pub/Sub 2. Store 3. Process

Slide 18

Slide 18 text

@ @gamussa @confluentinc Core abstraction

Slide 19

Slide 19 text

@ @gamussa @confluentinc Core abstraction DB - table

Slide 20

Slide 20 text

@ @gamussa @confluentinc Core abstraction DB - table Hadoop - file

Slide 21

Slide 21 text

@ @gamussa @confluentinc Core abstraction DB - table Hadoop - file Messaging -?

Slide 22

Slide 22 text

@ @gamussa @confluentinc LOGS

Slide 23

Slide 23 text

@ @gamussa @confluentinc Producing to Kafka Time

Slide 24

Slide 24 text

@ @gamussa @confluentinc Producing to Kafka Time C C C

Slide 25

Slide 25 text

@ @gamussa @confluentinc Producing to Kafka - With Key Time A B C D hash(key) % numPartitions = N

Slide 26

Slide 26 text

@ @gamussa @confluentinc Producing to Kafka - No Key Time Messages will be produced in a round robin fashion

Slide 27

Slide 27 text

@ @gamussa @confluentinc Producing to Kafka - No Key Time Messages will be produced in a round robin fashion

Slide 28

Slide 28 text

@ @gamussa @confluentinc Producing to Kafka - No Key Time Messages will be produced in a round robin fashion

Slide 29

Slide 29 text

@ @gamussa @confluentinc Producing to Kafka - No Key Time Messages will be produced in a round robin fashion

Slide 30

Slide 30 text

@ @gamussa @confluentinc Consuming From Kafka - Single Consumer C

Slide 31

Slide 31 text

@ @gamussa @confluentinc Consuming From Kafka - Grouped Consumers C C C1 C C C2

Slide 32

Slide 32 text

@ @gamussa @confluentinc Consuming From Kafka - Grouped Consumers C C C C

Slide 33

Slide 33 text

@ @gamussa @confluentinc Consuming From Kafka - Grouped Consumers 0 1 2 3

Slide 34

Slide 34 text

@ @gamussa @confluentinc Consuming From Kafka - Grouped Consumers 0 1 2 3

Slide 35

Slide 35 text

@ @gamussa @confluentinc Consuming From Kafka - Grouped Consumers 0, 3 1 2 3

Slide 36

Slide 36 text

@ @gamussa @confluentinc Producers Consumers

Slide 37

Slide 37 text

@ @gamussa @confluentinc

Slide 38

Slide 38 text

@ @gamussa @confluentinc

Slide 39

Slide 39 text

@ @gamussa @confluentinc

Slide 40

Slide 40 text

@ @gamussa @confluentinc Kafka Connect does hard work so you don’t

Slide 41

Slide 41 text

@ @gamussa @confluentinc Kafka Connect does hard work so you don’t 1. Scale out

Slide 42

Slide 42 text

@ @gamussa @confluentinc Kafka Connect does hard work so you don’t 1. Scale out

Slide 43

Slide 43 text

@ @gamussa @confluentinc Kafka Connect does hard work so you don’t 1. Scale out

Slide 44

Slide 44 text

@ @gamussa @confluentinc Kafka Connect does hard work so you don’t 1. Scale out

Slide 45

Slide 45 text

@ @gamussa @confluentinc

Slide 46

Slide 46 text

@ @gamussa @confluentinc

Slide 47

Slide 47 text

@ @gamussa @confluentinc

Slide 48

Slide 48 text

@ @gamussa @confluentinc

Slide 49

Slide 49 text

@ @gamussa @confluentinc Streaming Platform 1. Pub/Sub 2. Store 3. Process

Slide 50

Slide 50 text

@ @gamussa @confluentinc Why Store?

Slide 51

Slide 51 text

@ @gamussa @confluentinc Scalability of a filesystem

Slide 52

Slide 52 text

@ @gamussa @confluentinc Scalability of a filesystem Throughput 100s mb/s

Slide 53

Slide 53 text

@ @gamussa @confluentinc Scalability of a filesystem Throughput 100s mb/s TBs per server

Slide 54

Slide 54 text

@ @gamussa @confluentinc Scalability of a filesystem Throughput 100s mb/s TBs per server Commodity Hardware

Slide 55

Slide 55 text

@ @gamussa @confluentinc Scalability of a filesystem Throughput 100s mb/s TBs per server Commodity Hardware O(1) writes

Slide 56

Slide 56 text

@ @gamussa @confluentinc Guarantees of a database

Slide 57

Slide 57 text

@ @gamussa @confluentinc Guarantees of a database Persistence

Slide 58

Slide 58 text

@ @gamussa @confluentinc Guarantees of a database Persistence Strict ordering

Slide 59

Slide 59 text

@ @gamussa @confluentinc Distributed by Design

Slide 60

Slide 60 text

@ @gamussa @confluentinc Replication Distributed by Design

Slide 61

Slide 61 text

@ @gamussa @confluentinc Replication Fault Tolerance Distributed by Design

Slide 62

Slide 62 text

@ @gamussa @confluentinc Replication Fault Tolerance Partitioning Distributed by Design

Slide 63

Slide 63 text

@ @gamussa @confluentinc Replication Fault Tolerance Partitioning Scale Distributed by Design

Slide 64

Slide 64 text

@ @gamussa @confluentinc

Slide 65

Slide 65 text

@ @gamussa @confluentinc Partition Leadership and Replication Broker 1 Topic1 partition1 Broker 2 Broker 3 Broker 4 Topic1 partition1 Topic1 partition1 Leader Follower Topic1 partition2 Topic1 partition2 Topic1 partition2 Topic1 partition3 Topic1 partition4 Topic1 partition3 Topic1 partition3 Topic1 partition4 Topic1 partition4

Slide 66

Slide 66 text

@ @gamussa @confluentinc Partition Leadership and Replication - node failure Broker 1 Topic1 partition1 Broker 2 Broker 3 Broker 4 Topic1 partition1 Topic1 partition1 Leader Follower Topic1 partition2 Topic1 partition2 Topic1 partition2 Topic1 partition3 Topic1 partition4 Topic1 partition3 Topic1 partition3 Topic1 partition4 Topic1 partition4

Slide 67

Slide 67 text

@ @gamussa @confluentinc Streaming Platform 1. Pub/Sub 2. Store 3. Process

Slide 68

Slide 68 text

@ @gamussa @confluentinc What is Stream Processing? A machine for combining streams of events

Slide 69

Slide 69 text

@ @gamussa @confluentinc

Slide 70

Slide 70 text

@ @gamussa @confluentinc

Slide 71

Slide 71 text

@ @gamussa @confluentinc https://www.confluent.io/download/

Slide 72

Slide 72 text

@ @gamussa @confluentinc We are hiring! https://www.confluent.io/careers/

Slide 73

Slide 73 text

@ @gamussa @confluentinc One more thing…

Slide 74

Slide 74 text

@ @gamussa @confluentinc

Slide 75

Slide 75 text

@ @gamussa @confluentinc

Slide 76

Slide 76 text

@ @gamussa @confluentinc

Slide 77

Slide 77 text

@ @gamussa @confluentinc

Slide 78

Slide 78 text

@ @gamussa @confluentinc

Slide 79

Slide 79 text

@ @gamussa @confluentinc A Major New Paradigm

Slide 80

Slide 80 text

@ @gamussa @confluentinc Thanks! questions? @gamussa [email protected] We are hiring! https://www.confluent.io/careers/