Slide 1
Slide 1 text
A tale of queues
from ActiveMQ over Hazelcast to Disque
Philipp Krenn̴̴̴̴̴̴̴̴̴̴@xeraa
Slide 2
Slide 2 text
No content
Slide 3
Slide 3 text
No content
Slide 4
Slide 4 text
No content
Slide 5
Slide 5 text
Infrastructure | Developer Advocate
Slide 6
Slide 6 text
ViennaDB
Papers We Love Vienna
Slide 7
Slide 7 text
Electronic Data Interchange
Slide 8
Slide 8 text
No content
Slide 9
Slide 9 text
Automotive
Retail
Slide 10
Slide 10 text
Transmission & Message Flow
AS2, X.400, OFTP2, FTP, SMTP, IMAP,
web services,...
Slide 11
Slide 11 text
Document format
EDIFACT, X12, ODETTE, VDA,...
Slide 12
Slide 12 text
No content
Slide 13
Slide 13 text
No content
Slide 14
Slide 14 text
No content
Slide 15
Slide 15 text
No content
Slide 16
Slide 16 text
No content
Slide 17
Slide 17 text
Queues
Decoupling producer and consumer
Slide 18
Slide 18 text
!
Slide 19
Slide 19 text
!
Slide 20
Slide 20 text
Networked message queues like
ActiveMQ, RabbitMQ, ZeroMQ, and a
host of other Java inspired software
tumors are crutches of systems design.
— Ted Dziuba, http://widgetsandshit.com/teddziuba/2011/02/the-case-
against-queues.html
Slide 21
Slide 21 text
"The Blocking Consumer"
Slide 22
Slide 22 text
"Collecting Data for
Offline Processing"
Slide 23
Slide 23 text
"Everybody Loves System
Complexity"
Slide 24
Slide 24 text
There are only two hard problems in
distributed systems:
2. Exactly-once delivery
1. Guaranteed order of messages
2. Exactly-once delivery
— Mathias Verraes, https://twitter.com/mathiasverraes/status/632260618599403520
Slide 25
Slide 25 text
Order
As if it was on a single machine
http://book.mixu.net/distsys/time.html
Slide 26
Slide 26 text
http://blog.acolyer.org/2016/02/26/distributed-consistency-and-
session-anomalies/
Slide 27
Slide 27 text
Exactly once
Slide 28
Slide 28 text
Removal / ACK
Before processing: at-most-once
A!er processing: at-least-once
Slide 29
Slide 29 text
At-most-once
Requires strict consistency with
communication overhead
Network split — majority
Slide 30
Slide 30 text
At-least-once
Easier to distribute
Idempotent consumption
Timeout tuning
Slide 31
Slide 31 text
Queues and databases
Salvatore Sanfilippo 870 days ago, http://antirez.com/news/78
Slide 32
Slide 32 text
Cheating
Order
Slide 33
Slide 33 text
Cheating
1. Receive
2a. Persist in RDS (MySQL)
2b. Camel routing + queue
3. Acknowledge
Slide 34
Slide 34 text
Goals
Reliability
Liveness
At-most-once
Slide 35
Slide 35 text
Machines
Wait
Retry
Slide 36
Slide 36 text
Amazon Simple Queue
Service (SQS)
Fast, reliable, scalable, and fully
managed
Slide 37
Slide 37 text
No content
Slide 38
Slide 38 text
But
At-least-once
64KB (now 256KB and 2GB on S3)
payload
Slide 39
Slide 39 text
No content
Slide 40
Slide 40 text
Master Slave
Shared file system
JDBC
Replicated LevelDB
+ ZooKeeper
Slide 41
Slide 41 text
true
10000
10000
true
Slide 42
Slide 42 text
No content
Slide 43
Slide 43 text
RDS down !
Slide 44
Slide 44 text
No content
Slide 45
Slide 45 text
No content
Slide 46
Slide 46 text
In-memory data grid
Set, list, map, queue, topic, lock,
atomic long,...
Query, aggregate, MapReduce
Hibernate 2nd level cache, session
replication
Slide 47
Slide 47 text
Eierlegende
Wollmilchsau
Slide 48
Slide 48 text
"egg-laying wool-
milk-sow"
Slide 49
Slide 49 text
Great
Does everything for Java apps
Multi-master HA
Fast
Slide 50
Slide 50 text
Bad feeling
Very few queue users
Bugs with IP / DNS changes
Painful upgrades
Slide 51
Slide 51 text
Currently our action is to create
bug for these scenarios and try to
ensure exactly-once.
— Enes Akar, https://groups.google.com/forum/#!msg/hazelcast/
u_KLHVnvT_U/Qx5Km8COk_oJ
Slide 52
Slide 52 text
Kafka
Real-time focus
At-least-once
https://cwiki.apache.org/confluence/display/KAFKA/
Idempotent+Producer
Slide 53
Slide 53 text
Kestrel
Dead !
Slide 54
Slide 54 text
RabbitMQ
Too much middleware
Slide 55
Slide 55 text
In the event of network failure
(or a node crashing), messages
can be duplicated, and consumers
must be prepared to handle them.
— https://www.rabbitmq.com/reliability.html
Slide 56
Slide 56 text
No content
Slide 57
Slide 57 text
[...] no one try to use N Redis
independent nodes and the
offered primitives as a building
block for a distributed system [...]
— Salvatore Sanfilippo, http://antirez.com/news/78
Slide 58
Slide 58 text
Disque, an in-memory,
distributed job queue
— https://github.com/antirez/disque
Slide 59
Slide 59 text
Great
Does one thing
Multi-master HA
Configurable ACK
Similar to Redis
Slide 60
Slide 60 text
Driver: Spinach
"A scalable Java Disque client" based
on Lettuce, "A scalable Java Redis
client"
Slide 61
Slide 61 text
Mark Paluch
https://github.com/mp911de
Slide 62
Slide 62 text
!
Slide 63
Slide 63 text
Current configuration
Per message
Replication factor: 1
Retry time: 0
No persistence
Slide 64
Slide 64 text
WARNING: This is beta code and
may not be suitable for
production usage.
— https://github.com/antirez/disque
Slide 65
Slide 65 text
Disque 1.0 RC1 is out!
— Salvatore Sanfilippo 333 days ago, http://antirez.com/news/100
Slide 66
Slide 66 text
No content
Slide 67
Slide 67 text
No content
Slide 68
Slide 68 text
Back to Redis?
Slide 69
Slide 69 text
Conclusion
Slide 70
Slide 70 text
No content
Slide 71
Slide 71 text
No content
Slide 72
Slide 72 text
No content
Slide 73
Slide 73 text
No content
Slide 74
Slide 74 text
No content
Slide 75
Slide 75 text
Options
Disque (Redis)
ActiveMQ
Kafka
RabbitMQ
Hazelcast
SQS
Slide 76
Slide 76 text
Thank you!
Questions?!
@xeraa
Slide 77
Slide 77 text
Image Credit
· Schnitzel https://flic.kr/p/9m27wm
· Architektur https://flic.kr/p/6dwCAe
· Conchita https://flic.kr/p/nBqSHT
· Papier https://flic.kr/p/83thLf
· Lu!brücke https://commons.wikimedia.org/wiki/File:C-54landingattemplehof.jpg
· Bier https://flic.kr/p/aBSmtY
· Gum https://flic.kr/p/ALQ3b
· Eierlegende Wollmilchsau https://flic.kr/p/GzQTT
· Li! https://flic.kr/p/iBNr9x