@Apache KafkaA Streaming Data Platform
View Slide
@@gamussa @confluentincSolutions ArchitectDeveloper Advocate@gamussa in internetzHey you, yes, you,go follow me in twitter ©Who am I?
@@gamussa @confluentincA company is build onDATA FLOWSbutAll we have isDATA STORES
@@gamussa @confluentinc
@@gamussa @confluentincStreaming Platform1. Pub/Sub2. Store3. Process
@@gamussa @confluentincCore abstractionDB - tableHadoop - fileMessaging -?
@@gamussa @confluentincLOGS
@@gamussa @confluentincProducing to KafkaTime
@@gamussa @confluentincProducing to KafkaTimeC CC
@@gamussa @confluentincProducing to Kafka - With KeyTimeABCDhash(key) %numPartitions = N
@@gamussa @confluentincProducing to Kafka - No KeyTimeMessages will be produced in around robin fashion
@@gamussa @confluentincConsuming From Kafka - Single ConsumerC
@@gamussa @confluentincConsuming From Kafka - Grouped ConsumersCCC1CCC2
@@gamussa @confluentincConsuming From Kafka - Grouped ConsumersC CC C
@@gamussa @confluentincConsuming From Kafka - Grouped Consumers0 12 3
@@gamussa @confluentincConsuming From Kafka - Grouped Consumers0, 3 12 3
@@gamussa @confluentincProducers Consumers
@@gamussa @confluentincKafka Connect does hard work so you don’t1. Scale out
@@gamussa @confluentincWhyStore?
@@gamussa @confluentincScalability of a filesystemThroughput 100s mb/sTBs per serverCommodity HardwareO(1) writes
@@gamussa @confluentincGuarantees of a databasePersistenceStrict ordering
@@gamussa @confluentincReplicationFault TolerancePartitioningScaleDistributed by Design
@@gamussa @confluentincPartition Leadership and ReplicationBroker 1Topic1partition1Broker 2 Broker 3 Broker 4Topic1partition1Topic1partition1Leader FollowerTopic1partition2Topic1partition2Topic1partition2Topic1partition3Topic1partition4Topic1partition3Topic1partition3Topic1partition4Topic1partition4
@@gamussa @confluentincPartition Leadership and Replication - node failureBroker 1Topic1partition1Broker 2 Broker 3 Broker 4Topic1partition1Topic1partition1Leader FollowerTopic1partition2Topic1partition2Topic1partition2Topic1partition3Topic1partition4Topic1partition3Topic1partition3Topic1partition4Topic1partition4
@@gamussa @confluentincWhat is Stream Processing?A machine for combining streams of events
@@gamussa @confluentinchttps://www.confluent.io/download/
@@gamussa @confluentincWe are hiring!https://www.confluent.io/careers/
@@gamussa @confluentincOne more thing…
@@gamussa @confluentincA Major New Paradigm
@@gamussa @confluentincThanks!questions?@gamussa[email protected]We are hiring!https://www.confluent.io/careers/
vikgamov
miu_crescent
nagiss
bengo4com
mashiike
andpad
koudaiii
mploed
_kensh
miyashino
konifar
soudai
keathley
thekraken
davidbonilla
chriscoyier
mthomps
marcelosomers
malarkey
dougneiner
jponch
marktimemedia
jlugia
addyosmani