Scaling Twitter

Blaine
Big Bird.
(scaling twitter)
Rails Scales.
(but not out of the box)
First, Some Facts
• 600 requests per second. Growing fast.
• 180 Rails Instances (Mongrel). Growing fast.
• 1 Database Server (MySQL) + 1 Slave.
• 30-odd Processes for Misc. Jobs
• 8 Sun X4100s
• Many users, many updates.
Scaling Twitter
Scaling Twitter
Scaling Twitter
Joy          Pain




Oct   Nov   Dec    Jan   Feb     March   Apr
IM IN UR RAILZ




     MAKIN EM GO FAST
It’s Easy, Really.
1. Realize Your Site is Slow
2. Optimize the Database
3. Cache the Hell out of Everything
4. Scale Messaging
5. Deal With Abuse
It’s Easy, Really.
1. Realize Your Site is Slow
2. Optimize the Database
3. Cache the Hell out of Everything
4. Scale Messaging
5. Deal With Abuse
6. Profit
the
     more
      you
        know

{ Part the First }
We Failed at This.
Don’t Be Like Us

• Munin
• Nagios
• AWStats & Google Analytics
• Exception Notifier / Exception Logger
• Immediately add reporting to track problems.
Test Everything

•   Start Before You Start

•   No Need To Be Fancy

•   Tests Will Save Your Life

•   Agile Becomes
    Important When Your
    Site Is Down
<!-- served to you through a copper wire by sampaati at 22 Apr
    15:02 in 343 ms (d 102 / r 217). thank you, come again. -->
 <!-- served to you through a copper wire by kolea.twitter.com at
22 Apr 15:02 in 235 ms (d 87 / r 130). thank you, come again. -->
 <!-- served to you through a copper wire by raven.twitter.com at
22 Apr 15:01 in 450 ms (d 96 / r 337). thank you, come again. -->



                  Benchmarks?
                       let your users do it.
 <!-- served to you through a copper wire by kolea.twitter.com at
22 Apr 15:00 in 409 ms (d 88 / r 307). thank you, come again. -->
  <!-- served to you through a copper wire by firebird at 22 Apr
   15:03 in 2094 ms (d 643 / r 1445). thank you, come again. -->
   <!-- served to you through a copper wire by quetzal at 22 Apr
     15:01 in 384 ms (d 70 / r 297). thank you, come again. -->
The Database
  { Part the Second }
“The Next Application I Build is Going
to Be Easily Partitionable” - S. Butterfield
“The Next Application I Build is Going
to Be Easily Partitionable” - S. Butterfield
“The Next Application I Build is Going
to Be Easily Partitionable” - S. Butterfield
Too Late.
Index Everything
class AddIndex < ActiveRecord::Migration
     def self.up
       add_index :users, :email
     end

     def self.down
       remove_index :users, :email
     end
   end


Repeat for any column that appears in a WHERE clause

             Rails won’t do this for you.
Denormalize A Lot
class DenormalizeFriendsIds < ActiveRecord::Migration
  def self.up
    add_column "users", "friends_ids", :text
  end

  def self.down
    remove_column "users", "friends_ids"
  end
end
class Friendship < ActiveRecord::Base
  belongs_to :user
  belongs_to :friend

 after_create :add_to_denormalized_friends
 after_destroy :remove_from_denormalized_friends

  def add_to_denormalized_friends
    user.friends_ids << friend.id
    user.friends_ids.uniq!
    user.save_without_validation
  end

  def remove_from_denormalized_friends
    user.friends_ids.delete(friend.id)
    user.save_without_validation
  end
end
Don’t be Stupid
bob.friends.map(&:email)
     Status.count()
“email like ‘%#{search}%’”
That’s where we are.
                  Seriously.
  If your Rails application is doing anything more
complex than that, you’re doing something wrong*.



        * or you observed the First Rule of Butterfield.
Partitioning Comes Later.
   (we’ll let you know how it goes)
The Cache
 { Part the Third }
MemCache
MemCache
MemCache
!
class Status < ActiveRecord::Base
  class << self
    def count_with_memcache(*args)
      return count_without_memcache unless args.empty?
      count = CACHE.get(“status_count”)
      if count.nil?
        count = count_without_memcache
        CACHE.set(“status_count”, count)
      end
      count
    end
    alias_method_chain :count, :memcache
  end
  after_create :increment_memcache_count
  after_destroy :decrement_memcache_count
  ...
end
class User < ActiveRecord::Base
  def friends_statuses
    ids = CACHE.get(“friends_statuses:#{id}”)
    Status.find(:all, :conditions => [“id IN (?)”, ids])
  end
end

class Status < ActiveRecord::Base
  after_create :update_caches
  def update_caches
    user.friends_ids.each do |friend_id|
      ids = CACHE.get(“friends_statuses:#{friend_id}”)
      ids.pop
      ids.unshift(id)
      CACHE.set(“friends_statuses:#{friend_id}”, ids)
    end
  end
end
The Future


            ve d
          ti r
         co
         Ac
           e
         R
90% API Requests
     Cache Them!
“There are only two hard things in CS:
 cache invalidation and naming things.”

             – Phil Karlton, via Tim Bray
Messaging
{ Part the Fourth }
You Already Knew All
That Other Stuff, Right?
Producer             Consumer
           Message
Producer             Consumer
           Queue
Producer             Consumer
DRb
• The Good:
 • Stupid Easy
 • Reasonably Fast
• The Bad:
 • Kinda Flaky
 • Zero Redundancy
 • Tightly Coupled
ejabberd


            Jabber Client
                (drb)




           Incoming         Outgoing
Presence
           Messages         Messages


              MySQL
Server
     DRb.start_service ‘druby://localhost:10000’, myobject




                         Client
myobject = DRbObject.new_with_uri(‘druby://localhost:10000’)
Rinda

• Shared Queue (TupleSpace)
• Built with DRb
• RingyDingy makes it stupid easy
• See Eric Hodel’s documentation
• O(N) for take(). Sigh.
Timestamp: 12/22/06 01:53:14 (4 months ago)
      Author: lattice
      Message: Fugly. Seriously. Fugly.




        SELECT * FROM messages WHERE
substring(truncate(id,0),-2,1) = #{@fugly_dist_idx}
It Scales.
(except it stopped on Tuesday)
Options

• ActiveMQ (Java)
• RabbitMQ (erlang)
• MySQL + Lightweight Locking
• Something Else?
erlang?


What are you doing?
 Stabbing my eyes out with a fork.
Starling

• Ruby, will be ported to something faster
• 4000 transactional msgs/s
• First pass written in 4 hours
• Speaks MemCache (set, get)
Use Messages to
Invalidate Cache
   (it’s really not that hard)
Abuse
{ Part the Fifth }
The Italians
9000 friends in 24 hours
        (doesn’t scale)
http://flickr.com/photos/heather/464504545/
http://flickr.com/photos/curiouskiwi/165229284/
http://flickr.com/photo_zoom.gne?id=42914103&size=l
http://flickr.com/photos/madstillz/354596905/
http://flickr.com/photos/laughingsquid/382242677/
http://flickr.com/photos/bng/46678227/
1 of 56

Recommended

How Shopify Scales Rails by
How Shopify Scales RailsHow Shopify Scales Rails
How Shopify Scales Railsjduff
20.9K views63 slides
Architecting for the Cloud using NetflixOSS - Codemash Workshop by
Architecting for the Cloud using NetflixOSS - Codemash WorkshopArchitecting for the Cloud using NetflixOSS - Codemash Workshop
Architecting for the Cloud using NetflixOSS - Codemash WorkshopSudhir Tonse
39.6K views86 slides
Scalability, Availability & Stability Patterns by
Scalability, Availability & Stability PatternsScalability, Availability & Stability Patterns
Scalability, Availability & Stability PatternsJonas Bonér
516.3K views196 slides
Introduction to the Disruptor by
Introduction to the DisruptorIntroduction to the Disruptor
Introduction to the DisruptorTrisha Gee
68.7K views48 slides
facebook architecture for 600M users by
facebook architecture for 600M usersfacebook architecture for 600M users
facebook architecture for 600M usersJongyoon Choi
72K views60 slides
Introduction to Redis by
Introduction to RedisIntroduction to Redis
Introduction to RedisDvir Volk
121K views24 slides

More Related Content

What's hot

Unique ID generation in distributed systems by
Unique ID generation in distributed systemsUnique ID generation in distributed systems
Unique ID generation in distributed systemsDave Gardner
79.5K views31 slides
Introduction to memcached by
Introduction to memcachedIntroduction to memcached
Introduction to memcachedJurriaan Persyn
70.9K views77 slides
Improving PySpark performance: Spark Performance Beyond the JVM by
Improving PySpark performance: Spark Performance Beyond the JVMImproving PySpark performance: Spark Performance Beyond the JVM
Improving PySpark performance: Spark Performance Beyond the JVMHolden Karau
4.6K views49 slides
Kafka Retry and DLQ by
Kafka Retry and DLQKafka Retry and DLQ
Kafka Retry and DLQGeorge Teo
6.3K views72 slides
Explore your prometheus data in grafana - Promcon 2018 by
Explore your prometheus data in grafana - Promcon 2018Explore your prometheus data in grafana - Promcon 2018
Explore your prometheus data in grafana - Promcon 2018Grafana Labs
1.6K views33 slides
Big Data in Real-Time at Twitter by
Big Data in Real-Time at TwitterBig Data in Real-Time at Twitter
Big Data in Real-Time at Twitternkallen
139.3K views71 slides

What's hot(20)

Unique ID generation in distributed systems by Dave Gardner
Unique ID generation in distributed systemsUnique ID generation in distributed systems
Unique ID generation in distributed systems
Dave Gardner79.5K views
Improving PySpark performance: Spark Performance Beyond the JVM by Holden Karau
Improving PySpark performance: Spark Performance Beyond the JVMImproving PySpark performance: Spark Performance Beyond the JVM
Improving PySpark performance: Spark Performance Beyond the JVM
Holden Karau4.6K views
Kafka Retry and DLQ by George Teo
Kafka Retry and DLQKafka Retry and DLQ
Kafka Retry and DLQ
George Teo6.3K views
Explore your prometheus data in grafana - Promcon 2018 by Grafana Labs
Explore your prometheus data in grafana - Promcon 2018Explore your prometheus data in grafana - Promcon 2018
Explore your prometheus data in grafana - Promcon 2018
Grafana Labs1.6K views
Big Data in Real-Time at Twitter by nkallen
Big Data in Real-Time at TwitterBig Data in Real-Time at Twitter
Big Data in Real-Time at Twitter
nkallen139.3K views
Etsy Activity Feeds Architecture by Dan McKinley
Etsy Activity Feeds ArchitectureEtsy Activity Feeds Architecture
Etsy Activity Feeds Architecture
Dan McKinley112.1K views
Kafka streams windowing behind the curtain by confluent
Kafka streams windowing behind the curtain Kafka streams windowing behind the curtain
Kafka streams windowing behind the curtain
confluent1.3K views
Cassandra Introduction & Features by DataStax Academy
Cassandra Introduction & FeaturesCassandra Introduction & Features
Cassandra Introduction & Features
DataStax Academy31.9K views
Introducing BinarySortedMultiMap - A new Flink state primitive to boost your ... by Flink Forward
Introducing BinarySortedMultiMap - A new Flink state primitive to boost your ...Introducing BinarySortedMultiMap - A new Flink state primitive to boost your ...
Introducing BinarySortedMultiMap - A new Flink state primitive to boost your ...
Flink Forward579 views
Choosing an HDFS data storage format- Avro vs. Parquet and more - StampedeCon... by StampedeCon
Choosing an HDFS data storage format- Avro vs. Parquet and more - StampedeCon...Choosing an HDFS data storage format- Avro vs. Parquet and more - StampedeCon...
Choosing an HDFS data storage format- Avro vs. Parquet and more - StampedeCon...
StampedeCon129.5K views
How to build massive service for advance by DaeMyung Kang
How to build massive service for advanceHow to build massive service for advance
How to build massive service for advance
DaeMyung Kang13.2K views
Hive + Tez: A Performance Deep Dive by DataWorks Summit
Hive + Tez: A Performance Deep DiveHive + Tez: A Performance Deep Dive
Hive + Tez: A Performance Deep Dive
DataWorks Summit57.6K views
Performance Optimizations in Apache Impala by Cloudera, Inc.
Performance Optimizations in Apache ImpalaPerformance Optimizations in Apache Impala
Performance Optimizations in Apache Impala
Cloudera, Inc.10.7K views
NiFi Best Practices for the Enterprise by Gregory Keys
NiFi Best Practices for the EnterpriseNiFi Best Practices for the Enterprise
NiFi Best Practices for the Enterprise
Gregory Keys3.5K views
BI, Reporting and Analytics on Apache Cassandra by Victor Coustenoble
BI, Reporting and Analytics on Apache CassandraBI, Reporting and Analytics on Apache Cassandra
BI, Reporting and Analytics on Apache Cassandra
Victor Coustenoble27K views
Hive Anatomy by nzhang
Hive AnatomyHive Anatomy
Hive Anatomy
nzhang17.5K views
[NDC18] 야생의 땅 듀랑고의 데이터 엔지니어링 이야기: 로그 시스템 구축 경험 공유 by Hyojun Jeon
[NDC18] 야생의 땅 듀랑고의 데이터 엔지니어링 이야기: 로그 시스템 구축 경험 공유[NDC18] 야생의 땅 듀랑고의 데이터 엔지니어링 이야기: 로그 시스템 구축 경험 공유
[NDC18] 야생의 땅 듀랑고의 데이터 엔지니어링 이야기: 로그 시스템 구축 경험 공유
Hyojun Jeon15.8K views
Kafka replication apachecon_2013 by Jun Rao
Kafka replication apachecon_2013Kafka replication apachecon_2013
Kafka replication apachecon_2013
Jun Rao21.3K views

Similar to Scaling Twitter

Hiveminder - Everything but the Secret Sauce by
Hiveminder - Everything but the Secret SauceHiveminder - Everything but the Secret Sauce
Hiveminder - Everything but the Secret SauceJesse Vincent
2.9K views265 slides
Beijing Perl Workshop 2008 Hiveminder Secret Sauce by
Beijing Perl Workshop 2008 Hiveminder Secret SauceBeijing Perl Workshop 2008 Hiveminder Secret Sauce
Beijing Perl Workshop 2008 Hiveminder Secret SauceJesse Vincent
1.2K views238 slides
Microblogging via XMPP by
Microblogging via XMPPMicroblogging via XMPP
Microblogging via XMPPStoyan Zhekov
2.5K views39 slides
Aprendendo solid com exemplos by
Aprendendo solid com exemplosAprendendo solid com exemplos
Aprendendo solid com exemplosvinibaggio
1.1K views52 slides
Socket applications by
Socket applicationsSocket applications
Socket applicationsJoão Moura
601 views121 slides
From crash to testcase by
From crash to testcaseFrom crash to testcase
From crash to testcaseRoel Van de Paar
1.6K views54 slides

Similar to Scaling Twitter(20)

Hiveminder - Everything but the Secret Sauce by Jesse Vincent
Hiveminder - Everything but the Secret SauceHiveminder - Everything but the Secret Sauce
Hiveminder - Everything but the Secret Sauce
Jesse Vincent2.9K views
Beijing Perl Workshop 2008 Hiveminder Secret Sauce by Jesse Vincent
Beijing Perl Workshop 2008 Hiveminder Secret SauceBeijing Perl Workshop 2008 Hiveminder Secret Sauce
Beijing Perl Workshop 2008 Hiveminder Secret Sauce
Jesse Vincent1.2K views
Microblogging via XMPP by Stoyan Zhekov
Microblogging via XMPPMicroblogging via XMPP
Microblogging via XMPP
Stoyan Zhekov2.5K views
Aprendendo solid com exemplos by vinibaggio
Aprendendo solid com exemplosAprendendo solid com exemplos
Aprendendo solid com exemplos
vinibaggio1.1K views
Socket applications by João Moura
Socket applicationsSocket applications
Socket applications
João Moura601 views
Dynomite at Erlang Factory by moonpolysoft
Dynomite at Erlang FactoryDynomite at Erlang Factory
Dynomite at Erlang Factory
moonpolysoft594 views
Performance Optimization of Rails Applications by Serge Smetana
Performance Optimization of Rails ApplicationsPerformance Optimization of Rails Applications
Performance Optimization of Rails Applications
Serge Smetana8.8K views
Ensuring High Availability for Real-time Analytics featuring Boxed Ice / Serv... by MongoDB
Ensuring High Availability for Real-time Analytics featuring Boxed Ice / Serv...Ensuring High Availability for Real-time Analytics featuring Boxed Ice / Serv...
Ensuring High Availability for Real-time Analytics featuring Boxed Ice / Serv...
MongoDB602 views
WebPerformance: Why and How? – Stefan Wintermeyer by Elixir Club
WebPerformance: Why and How? – Stefan WintermeyerWebPerformance: Why and How? – Stefan Wintermeyer
WebPerformance: Why and How? – Stefan Wintermeyer
Elixir Club576 views
NPW2009 - my.opera.com scalability v2.0 by Cosimo Streppone
NPW2009 - my.opera.com scalability v2.0NPW2009 - my.opera.com scalability v2.0
NPW2009 - my.opera.com scalability v2.0
Cosimo Streppone1K views
Fisl - Deployment by Fabio Akita
Fisl - DeploymentFisl - Deployment
Fisl - Deployment
Fabio Akita934 views
SD, a P2P bug tracking system by Jesse Vincent
SD, a P2P bug tracking systemSD, a P2P bug tracking system
SD, a P2P bug tracking system
Jesse Vincent13.6K views
RubyEnRails2007 - Dr Nic Williams - Keynote by Dr Nic Williams
RubyEnRails2007 - Dr Nic Williams - KeynoteRubyEnRails2007 - Dr Nic Williams - Keynote
RubyEnRails2007 - Dr Nic Williams - Keynote
Dr Nic Williams2.8K views
MongoDB: Optimising for Performance, Scale & Analytics by Server Density
MongoDB: Optimising for Performance, Scale & AnalyticsMongoDB: Optimising for Performance, Scale & Analytics
MongoDB: Optimising for Performance, Scale & Analytics
Server Density555 views
JDD2015: Sharding with Akka Cluster: From Theory to Production - Krzysztof Ot... by PROIDEA
JDD2015: Sharding with Akka Cluster: From Theory to Production - Krzysztof Ot...JDD2015: Sharding with Akka Cluster: From Theory to Production - Krzysztof Ot...
JDD2015: Sharding with Akka Cluster: From Theory to Production - Krzysztof Ot...
PROIDEA125 views
Web 2.0 Performance and Reliability: How to Run Large Web Apps by adunne
Web 2.0 Performance and Reliability: How to Run Large Web AppsWeb 2.0 Performance and Reliability: How to Run Large Web Apps
Web 2.0 Performance and Reliability: How to Run Large Web Apps
adunne2.4K views
How to avoid hanging yourself with Rails by Rowan Hick
How to avoid hanging yourself with RailsHow to avoid hanging yourself with Rails
How to avoid hanging yourself with Rails
Rowan Hick1.3K views
Monkeybars in the Manor by martinbtt
Monkeybars in the ManorMonkeybars in the Manor
Monkeybars in the Manor
martinbtt820 views

More from Blaine

Social Privacy for HTTP over Webfinger by
Social Privacy for HTTP over WebfingerSocial Privacy for HTTP over Webfinger
Social Privacy for HTTP over WebfingerBlaine
5.5K views23 slides
Social Software for Robots by
Social Software for RobotsSocial Software for Robots
Social Software for RobotsBlaine
1.4K views46 slides
OAuth by
OAuthOAuth
OAuthBlaine
4K views37 slides
Building the Real Time Web by
Building the Real Time WebBuilding the Real Time Web
Building the Real Time WebBlaine
2.6K views77 slides
You & Me & Everyone We Know by
You & Me & Everyone We KnowYou & Me & Everyone We Know
You & Me & Everyone We KnowBlaine
2.4K views54 slides
Social Software for Robots by
Social Software for RobotsSocial Software for Robots
Social Software for RobotsBlaine
14.2K views50 slides

More from Blaine(6)

Social Privacy for HTTP over Webfinger by Blaine
Social Privacy for HTTP over WebfingerSocial Privacy for HTTP over Webfinger
Social Privacy for HTTP over Webfinger
Blaine5.5K views
Social Software for Robots by Blaine
Social Software for RobotsSocial Software for Robots
Social Software for Robots
Blaine1.4K views
OAuth by Blaine
OAuthOAuth
OAuth
Blaine4K views
Building the Real Time Web by Blaine
Building the Real Time WebBuilding the Real Time Web
Building the Real Time Web
Blaine2.6K views
You & Me & Everyone We Know by Blaine
You & Me & Everyone We KnowYou & Me & Everyone We Know
You & Me & Everyone We Know
Blaine2.4K views
Social Software for Robots by Blaine
Social Software for RobotsSocial Software for Robots
Social Software for Robots
Blaine14.2K views

Recently uploaded

DRBD Deep Dive - Philipp Reisner - LINBIT by
DRBD Deep Dive - Philipp Reisner - LINBITDRBD Deep Dive - Philipp Reisner - LINBIT
DRBD Deep Dive - Philipp Reisner - LINBITShapeBlue
110 views21 slides
Migrating VMware Infra to KVM Using CloudStack - Nicolas Vazquez - ShapeBlue by
Migrating VMware Infra to KVM Using CloudStack - Nicolas Vazquez - ShapeBlueMigrating VMware Infra to KVM Using CloudStack - Nicolas Vazquez - ShapeBlue
Migrating VMware Infra to KVM Using CloudStack - Nicolas Vazquez - ShapeBlueShapeBlue
147 views20 slides
The Power of Heat Decarbonisation Plans in the Built Environment by
The Power of Heat Decarbonisation Plans in the Built EnvironmentThe Power of Heat Decarbonisation Plans in the Built Environment
The Power of Heat Decarbonisation Plans in the Built EnvironmentIES VE
67 views20 slides
Zero to Cloud Hero: Crafting a Private Cloud from Scratch with XCP-ng, Xen Or... by
Zero to Cloud Hero: Crafting a Private Cloud from Scratch with XCP-ng, Xen Or...Zero to Cloud Hero: Crafting a Private Cloud from Scratch with XCP-ng, Xen Or...
Zero to Cloud Hero: Crafting a Private Cloud from Scratch with XCP-ng, Xen Or...ShapeBlue
128 views20 slides
Declarative Kubernetes Cluster Deployment with Cloudstack and Cluster API - O... by
Declarative Kubernetes Cluster Deployment with Cloudstack and Cluster API - O...Declarative Kubernetes Cluster Deployment with Cloudstack and Cluster API - O...
Declarative Kubernetes Cluster Deployment with Cloudstack and Cluster API - O...ShapeBlue
59 views13 slides
"Surviving highload with Node.js", Andrii Shumada by
"Surviving highload with Node.js", Andrii Shumada "Surviving highload with Node.js", Andrii Shumada
"Surviving highload with Node.js", Andrii Shumada Fwdays
49 views29 slides

Recently uploaded(20)

DRBD Deep Dive - Philipp Reisner - LINBIT by ShapeBlue
DRBD Deep Dive - Philipp Reisner - LINBITDRBD Deep Dive - Philipp Reisner - LINBIT
DRBD Deep Dive - Philipp Reisner - LINBIT
ShapeBlue110 views
Migrating VMware Infra to KVM Using CloudStack - Nicolas Vazquez - ShapeBlue by ShapeBlue
Migrating VMware Infra to KVM Using CloudStack - Nicolas Vazquez - ShapeBlueMigrating VMware Infra to KVM Using CloudStack - Nicolas Vazquez - ShapeBlue
Migrating VMware Infra to KVM Using CloudStack - Nicolas Vazquez - ShapeBlue
ShapeBlue147 views
The Power of Heat Decarbonisation Plans in the Built Environment by IES VE
The Power of Heat Decarbonisation Plans in the Built EnvironmentThe Power of Heat Decarbonisation Plans in the Built Environment
The Power of Heat Decarbonisation Plans in the Built Environment
IES VE67 views
Zero to Cloud Hero: Crafting a Private Cloud from Scratch with XCP-ng, Xen Or... by ShapeBlue
Zero to Cloud Hero: Crafting a Private Cloud from Scratch with XCP-ng, Xen Or...Zero to Cloud Hero: Crafting a Private Cloud from Scratch with XCP-ng, Xen Or...
Zero to Cloud Hero: Crafting a Private Cloud from Scratch with XCP-ng, Xen Or...
ShapeBlue128 views
Declarative Kubernetes Cluster Deployment with Cloudstack and Cluster API - O... by ShapeBlue
Declarative Kubernetes Cluster Deployment with Cloudstack and Cluster API - O...Declarative Kubernetes Cluster Deployment with Cloudstack and Cluster API - O...
Declarative Kubernetes Cluster Deployment with Cloudstack and Cluster API - O...
ShapeBlue59 views
"Surviving highload with Node.js", Andrii Shumada by Fwdays
"Surviving highload with Node.js", Andrii Shumada "Surviving highload with Node.js", Andrii Shumada
"Surviving highload with Node.js", Andrii Shumada
Fwdays49 views
Elevating Privacy and Security in CloudStack - Boris Stoyanov - ShapeBlue by ShapeBlue
Elevating Privacy and Security in CloudStack - Boris Stoyanov - ShapeBlueElevating Privacy and Security in CloudStack - Boris Stoyanov - ShapeBlue
Elevating Privacy and Security in CloudStack - Boris Stoyanov - ShapeBlue
ShapeBlue149 views
How to Re-use Old Hardware with CloudStack. Saving Money and the Environment ... by ShapeBlue
How to Re-use Old Hardware with CloudStack. Saving Money and the Environment ...How to Re-use Old Hardware with CloudStack. Saving Money and the Environment ...
How to Re-use Old Hardware with CloudStack. Saving Money and the Environment ...
ShapeBlue97 views
KVM Security Groups Under the Hood - Wido den Hollander - Your.Online by ShapeBlue
KVM Security Groups Under the Hood - Wido den Hollander - Your.OnlineKVM Security Groups Under the Hood - Wido den Hollander - Your.Online
KVM Security Groups Under the Hood - Wido den Hollander - Your.Online
ShapeBlue154 views
NTGapps NTG LowCode Platform by Mustafa Kuğu
NTGapps NTG LowCode Platform NTGapps NTG LowCode Platform
NTGapps NTG LowCode Platform
Mustafa Kuğu287 views
Centralized Logging Feature in CloudStack using ELK and Grafana - Kiran Chava... by ShapeBlue
Centralized Logging Feature in CloudStack using ELK and Grafana - Kiran Chava...Centralized Logging Feature in CloudStack using ELK and Grafana - Kiran Chava...
Centralized Logging Feature in CloudStack using ELK and Grafana - Kiran Chava...
ShapeBlue74 views
Data Integrity for Banking and Financial Services by Precisely
Data Integrity for Banking and Financial ServicesData Integrity for Banking and Financial Services
Data Integrity for Banking and Financial Services
Precisely76 views
Keynote Talk: Open Source is Not Dead - Charles Schulz - Vates by ShapeBlue
Keynote Talk: Open Source is Not Dead - Charles Schulz - VatesKeynote Talk: Open Source is Not Dead - Charles Schulz - Vates
Keynote Talk: Open Source is Not Dead - Charles Schulz - Vates
ShapeBlue178 views
Igniting Next Level Productivity with AI-Infused Data Integration Workflows by Safe Software
Igniting Next Level Productivity with AI-Infused Data Integration Workflows Igniting Next Level Productivity with AI-Infused Data Integration Workflows
Igniting Next Level Productivity with AI-Infused Data Integration Workflows
Safe Software373 views
CloudStack and GitOps at Enterprise Scale - Alex Dometrius, Rene Glover - AT&T by ShapeBlue
CloudStack and GitOps at Enterprise Scale - Alex Dometrius, Rene Glover - AT&TCloudStack and GitOps at Enterprise Scale - Alex Dometrius, Rene Glover - AT&T
CloudStack and GitOps at Enterprise Scale - Alex Dometrius, Rene Glover - AT&T
ShapeBlue81 views
Developments to CloudStack’s SDN ecosystem: Integration with VMWare NSX 4 - P... by ShapeBlue
Developments to CloudStack’s SDN ecosystem: Integration with VMWare NSX 4 - P...Developments to CloudStack’s SDN ecosystem: Integration with VMWare NSX 4 - P...
Developments to CloudStack’s SDN ecosystem: Integration with VMWare NSX 4 - P...
ShapeBlue120 views
Confidence in CloudStack - Aron Wagner, Nathan Gleason - Americ by ShapeBlue
Confidence in CloudStack - Aron Wagner, Nathan Gleason - AmericConfidence in CloudStack - Aron Wagner, Nathan Gleason - Americ
Confidence in CloudStack - Aron Wagner, Nathan Gleason - Americ
ShapeBlue58 views
Backroll, News and Demo - Pierre Charton, Matthias Dhellin, Ousmane Diarra - ... by ShapeBlue
Backroll, News and Demo - Pierre Charton, Matthias Dhellin, Ousmane Diarra - ...Backroll, News and Demo - Pierre Charton, Matthias Dhellin, Ousmane Diarra - ...
Backroll, News and Demo - Pierre Charton, Matthias Dhellin, Ousmane Diarra - ...
ShapeBlue121 views

Scaling Twitter

  • 2. Rails Scales. (but not out of the box)
  • 3. First, Some Facts • 600 requests per second. Growing fast. • 180 Rails Instances (Mongrel). Growing fast. • 1 Database Server (MySQL) + 1 Slave. • 30-odd Processes for Misc. Jobs • 8 Sun X4100s • Many users, many updates.
  • 7. Joy Pain Oct Nov Dec Jan Feb March Apr
  • 8. IM IN UR RAILZ MAKIN EM GO FAST
  • 9. It’s Easy, Really. 1. Realize Your Site is Slow 2. Optimize the Database 3. Cache the Hell out of Everything 4. Scale Messaging 5. Deal With Abuse
  • 10. It’s Easy, Really. 1. Realize Your Site is Slow 2. Optimize the Database 3. Cache the Hell out of Everything 4. Scale Messaging 5. Deal With Abuse 6. Profit
  • 11. the more you know { Part the First }
  • 12. We Failed at This.
  • 13. Don’t Be Like Us • Munin • Nagios • AWStats & Google Analytics • Exception Notifier / Exception Logger • Immediately add reporting to track problems.
  • 14. Test Everything • Start Before You Start • No Need To Be Fancy • Tests Will Save Your Life • Agile Becomes Important When Your Site Is Down
  • 15. <!-- served to you through a copper wire by sampaati at 22 Apr 15:02 in 343 ms (d 102 / r 217). thank you, come again. --> <!-- served to you through a copper wire by kolea.twitter.com at 22 Apr 15:02 in 235 ms (d 87 / r 130). thank you, come again. --> <!-- served to you through a copper wire by raven.twitter.com at 22 Apr 15:01 in 450 ms (d 96 / r 337). thank you, come again. --> Benchmarks? let your users do it. <!-- served to you through a copper wire by kolea.twitter.com at 22 Apr 15:00 in 409 ms (d 88 / r 307). thank you, come again. --> <!-- served to you through a copper wire by firebird at 22 Apr 15:03 in 2094 ms (d 643 / r 1445). thank you, come again. --> <!-- served to you through a copper wire by quetzal at 22 Apr 15:01 in 384 ms (d 70 / r 297). thank you, come again. -->
  • 16. The Database { Part the Second }
  • 17. “The Next Application I Build is Going to Be Easily Partitionable” - S. Butterfield
  • 18. “The Next Application I Build is Going to Be Easily Partitionable” - S. Butterfield
  • 19. “The Next Application I Build is Going to Be Easily Partitionable” - S. Butterfield
  • 22. class AddIndex < ActiveRecord::Migration def self.up add_index :users, :email end def self.down remove_index :users, :email end end Repeat for any column that appears in a WHERE clause Rails won’t do this for you.
  • 24. class DenormalizeFriendsIds < ActiveRecord::Migration def self.up add_column "users", "friends_ids", :text end def self.down remove_column "users", "friends_ids" end end
  • 25. class Friendship < ActiveRecord::Base belongs_to :user belongs_to :friend after_create :add_to_denormalized_friends after_destroy :remove_from_denormalized_friends def add_to_denormalized_friends user.friends_ids << friend.id user.friends_ids.uniq! user.save_without_validation end def remove_from_denormalized_friends user.friends_ids.delete(friend.id) user.save_without_validation end end
  • 27. bob.friends.map(&:email) Status.count() “email like ‘%#{search}%’”
  • 28. That’s where we are. Seriously. If your Rails application is doing anything more complex than that, you’re doing something wrong*. * or you observed the First Rule of Butterfield.
  • 29. Partitioning Comes Later. (we’ll let you know how it goes)
  • 30. The Cache { Part the Third }
  • 34. !
  • 35. class Status < ActiveRecord::Base class << self def count_with_memcache(*args) return count_without_memcache unless args.empty? count = CACHE.get(“status_count”) if count.nil? count = count_without_memcache CACHE.set(“status_count”, count) end count end alias_method_chain :count, :memcache end after_create :increment_memcache_count after_destroy :decrement_memcache_count ... end
  • 36. class User < ActiveRecord::Base def friends_statuses ids = CACHE.get(“friends_statuses:#{id}”) Status.find(:all, :conditions => [“id IN (?)”, ids]) end end class Status < ActiveRecord::Base after_create :update_caches def update_caches user.friends_ids.each do |friend_id| ids = CACHE.get(“friends_statuses:#{friend_id}”) ids.pop ids.unshift(id) CACHE.set(“friends_statuses:#{friend_id}”, ids) end end end
  • 37. The Future ve d ti r co Ac e R
  • 38. 90% API Requests Cache Them!
  • 39. “There are only two hard things in CS: cache invalidation and naming things.” – Phil Karlton, via Tim Bray
  • 41. You Already Knew All That Other Stuff, Right?
  • 42. Producer Consumer Message Producer Consumer Queue Producer Consumer
  • 43. DRb • The Good: • Stupid Easy • Reasonably Fast • The Bad: • Kinda Flaky • Zero Redundancy • Tightly Coupled
  • 44. ejabberd Jabber Client (drb) Incoming Outgoing Presence Messages Messages MySQL
  • 45. Server DRb.start_service ‘druby://localhost:10000’, myobject Client myobject = DRbObject.new_with_uri(‘druby://localhost:10000’)
  • 46. Rinda • Shared Queue (TupleSpace) • Built with DRb • RingyDingy makes it stupid easy • See Eric Hodel’s documentation • O(N) for take(). Sigh.
  • 47. Timestamp: 12/22/06 01:53:14 (4 months ago) Author: lattice Message: Fugly. Seriously. Fugly. SELECT * FROM messages WHERE substring(truncate(id,0),-2,1) = #{@fugly_dist_idx}
  • 48. It Scales. (except it stopped on Tuesday)
  • 49. Options • ActiveMQ (Java) • RabbitMQ (erlang) • MySQL + Lightweight Locking • Something Else?
  • 50. erlang? What are you doing? Stabbing my eyes out with a fork.
  • 51. Starling • Ruby, will be ported to something faster • 4000 transactional msgs/s • First pass written in 4 hours • Speaks MemCache (set, get)
  • 52. Use Messages to Invalidate Cache (it’s really not that hard)
  • 55. 9000 friends in 24 hours (doesn’t scale)