Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Sign up for free
Menu
Search
Features
All features
Private URLs
Password Protection
Custom URLS
Scheduled publishing
Remove Branding
Restrict embedding
Deck Collections
Notes
Features
All features
Private URLs
Password Protection
Custom URLS
Scheduled publishing
Remove Branding
Restrict embedding
Deck Collections
Notes
Explore
Featured decks
Featured speakers
Programming
Technology
Storyboards
Explore
Featured decks
Featured speakers
Programming
Technology
Storyboards
Pricing
Search
Sign in
Sign up for free
Kafka Will Get The Message Across, Guaranteed.
Search
David Zuelke
January 28, 2017
Programming
340
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
Kafka Will Get The Message Across, Guaranteed.
Presentation given at PHP Benelux 2017 near Antwerp, Belgium.
David Zuelke
January 28, 2017
Transcript
84 slides
KAFKA WILL GET THE MESSAGE ACROSS. GUARANTEED. PHP Benelux 2017
Belgium
David Zuelke
None
[email protected]
@dzuelke
KAFKA
LinkedIn
APACHE KAFKA
"uh oh, another Apache project?!"
None
KEEP CALM AND LOOK AT THE WEBSITE
None
"Basically it is a massively scalable pub/sub message queue. architected
as a distributed transaction log."
"so it's a queue?"
it's not just a queue
queues are not multi-subscriber :(
"so it's a pubsub thing?"
it's not just a pubsub thing
pubsub broadcasts to all subscribers :(
it's a log
None
not that kind of log
WAL
Write-Ahead Log
WRITE-AHEAD LOG
None
1 foo 2 bar 3 baz 4 hi
1 create document: "foo", data: "…" 2 update document: "foo",
data: "…" 3 create document: "bar", data: "…" 4 remove document: "foo"
None
never corrupts
sequential I/O
None
sequential I/O
every message will be read at least once, no random
access
FileChannel.transferTo (shovels data straight from e.g. disk cache to network
interface, no copying via RAM)
"HI, I AM KAFKA" "Buckle up while we process (m|b|tr)illions
of messages/s."
TOPICS
streams of records
1 2 3 4 5 6 7 …
1 2 3 4 5 6 7 8 … producer
writes consumer reads
can have many subscribers
1 2 3 4 5 6 7 8 … producer
writes consumerB reads consumerA reads
can be partitioned
P0 1 2 3 4 5 6 7 … P1
1 2 3 4 … P2 1 2 3 4 5 6 7 8 … P3 1 2 3 4 5 6 …
partitions let you scale storage!
partitions let you scale consuming!
None
all records are retained, whether consumed or not, up to
a configurable limit
PRODUCERS
byte[]
(typically JSON, XML, Avro, Thrift, Protobufs)
(typically not funny GIFs)
can choose explicit partition, or a key (which is used
for auto-partitioning)
https://github.com/edenhill/librdkafka & https://arnaud-lb.github.io/php-rdkafka/
BASIC PRODUCER $rk = new RdKafka\Producer(); $rk->addBrokers("127.0.0.1"); $topic = $rk->newTopic("test");
$topic->produce(RD_KAFKA_PARTITION_UA, 0, "Unassigned partition, let Kafka choose"); $topic->produce(RD_KAFKA_PARTITION_UA, 0, "Yay consistent hashing", $user->getId()); $topic->produce(1, 0, "This will always be sent to partition 1");
CONSUMERS
cheap
only metadata stored per consumer: offset
guaranteed to always have messages in right order (within a
partition)
can themselves produce new messages! (but there is also a
Streams API for pure transformations)
None
BASIC CONSUMER $conf = new RdKafka\Conf(); $conf->set('group.id', 'myConsumerGroup'); $rk =
new RdKafka\Consumer($conf); $rk->addBrokers("127.0.0.1"); $topicConf = new RdKafka\TopicConf(); $topicConf->set('auto.commit.interval.ms', 100); $topic = $rk->newTopic("test", $topicConf); $topic->consumeStart(0, RD_KAFKA_OFFSET_STORED); while (true) { $msg = $topic->consume(0, 120*10000); do_something($msg); }
AT-MOST ONCE DELIVERY $conf = new RdKafka\Conf(); $conf->set('group.id', 'myConsumerGroup'); $rk
= new RdKafka\Consumer($conf); $rk->addBrokers("127.0.0.1"); $topicConf = new RdKafka\TopicConf(); $topicConf->set('auto.commit.enable', false); $topic = $rk->newTopic("test", $topicConf); $topic->consumeStart(0, RD_KAFKA_OFFSET_STORED); while (true) { $msg = $topic->consume(0, 120*10000); $topic->offsetStore($msg->partition, $msg->offset); do_something($msg); }
AT-LEAST ONCE DELIVERY $conf = new RdKafka\Conf(); $conf->set('group.id', 'myConsumerGroup'); $rk
= new RdKafka\Consumer($conf); $rk->addBrokers("127.0.0.1"); $topicConf = new RdKafka\TopicConf(); $topicConf->set('auto.commit.enable', false); $topic = $rk->newTopic("test", $topicConf); $topic->consumeStart(0, RD_KAFKA_OFFSET_STORED); while (true) { $msg = $topic->consume(0, 120*10000); do_something($msg); $topic->offsetStore($msg->partition, $msg->offset); }
EXACTLY-ONCE DELIVERY
you cannot have exactly-once delivery
THE BYZANTINE GENERALS "together we can beat the monsters. let's
both attack at 07:00?" "confirm, we attack at 07:00" ☠
USE CASES
• LinkedIn • Yahoo • Twitter • Netflix • Square
• Spotify • Pinterest • Uber • Goldman Sachs • Tumblr • PayPal • Airbnb • Mozilla • Cisco • Etsy • Foursquare • Shopify • CloudFlare
ingest the Twitter firehose and turn it into a pointless
demo ;)
None
messaging, of course
track user activity
record runtime metrics
aggregate logs
IoT (you could still e.g. use MQTT over the wire,
and bridge to Kafka)
replicate information between data centers (also see Connector API)
Event Sourcing broker :)
WAL / Commit Log for another system
billing!
"shock absorber" between systems to avoid overload of DBs, APIs,
etc.
in PHP: mostly producing messages; better languages exist for consuming
The End
THANK YOU FOR LISTENING! Questions? Ask me: @dzuelke &
[email protected]
More Decks by David Zuelke
See All by David Zuelke
Your next Web server will be written in... PHP
dzuelke
0
190
Getting Things Done
dzuelke
1
500
Your next Web server will be written in... PHP
dzuelke
2
330
Your next Web server will be written in... PHP
dzuelke
3
1.3k
Kafka Will Get The Message Across, Guaranteed.
dzuelke
0
930
Heroku at BattleHack Venice 2015
dzuelke
0
160
Designing HTTP Interfaces and RESTful Web Services
dzuelke
6
1.6k
The Twelve-Factor App: Best Practices for Modern Web Applications
dzuelke
4
590
Designing HTTP Interfaces and RESTful Web Services
dzuelke
6
560
Other Decks in Programming
See All in Programming
Materialが消えた日 〜真っ白なFlutterキャンバスにRenderObjectで絵を描く〜
umigishiaoi
0
170
Getting Started with Vibe Coding
miccheng
0
240
AI時代のシニアエンジニアとは
yosuke_furukawa
PRO
15
12k
protojson における文字化けの扱い
kazuki1023
0
450
作れるエンジニアから、価値を届けるエンジニアへ
ryunakayama
1
320
コード生成時代のAPI定義を支えるRequest Spec移行
pvcresin
0
290
私なりのハーネスエンジニアリング ~独自の言語処理系の開発を通して~
seosoft
0
230
GraphQL の onError と Nullability のこれから
quramy
0
220
テストだけで見つかる? Rubyでつくるメソッド発掘器
izawa
0
150
Agentic AI with the Power of Jakarta EE
ivargrimstad
0
100
Crash Auto Pipelines
veronikapj
1
440
[Kotlin Fest 2026] eBPF Tracing for Kotlin
takehaya
1
390
Featured
See All Featured
ピンチをチャンスに:未来をつくるプロダクトロードマップ #pmconf2020
aki_iinuma
128
56k
HDC tutorial
michielstock
2
1k
Facilitating Awesome Meetings
lara
57
7.2k
4 Signs Your Business is Dying
shpigford
187
23k
Designing for Timeless Needs
cassininazir
1
570
Music & Morning Musume
bryan
48
7.5k
YesSQL, Process and Tooling at Scale
rocio
174
15k
Marketing to machines
jonoalderson
1
6k
Navigating the moral maze — ethical principles for Al-driven product design
skipperchong
2
690
ReactJS: Keep Simple. Everything can be a component!
pedronauck
666
130k
The AI Revolution Will Not Be Monopolized: How open-source beats economies of scale, even for LLMs
inesmontani
PRO
3
3.8k
The Hidden Cost of Media on the Web [PixelPalooza 2025]
tammyeverts
2
580