Webinar: Scaling MySQL
Catch 22 of Read Write Splitting
                                   January 17, 2013
Agenda


       1. Who We Are

       2. The Scalability Problem

       3. How We Solve it with Read/Write Splitting

       4. Customer ROI/Case Studies

       5. Q & A
          (please type questions directly into the GoToWebinar side panel)




2
Who We Are

    Presenters:                                     Paul Campaniello,
                                                  VP of Global Marketing
                                              25 year technology veteran with
                                              marketing experience at Mendix,
                                              Lumigent, Savantis and Precise.




                Doron Levari, Founder
            A technologist and long-time
          veteran of the database industry.
         Prior to founding ScaleBase, Doron
                  was CEO to Aluna.


3
Who We Are


ScaleBase allows apps
to cost-effectively scale
to an infinite number of users,
with NO disruption to the existing infrastructure




4
The ScaleBase Data Traffic Manager

                                                     •    Database Scalability
                                                           –   Scale out relational databases
                                                               to unlimited users
                                                           –   Real-time elasticity

                                                     •    Database Availability
                                                           –   Enable high availability of all
                                                               apps

                                                     •    Centralized Management
                                                           –   Removes complexity and
                                                               provides a unified point of
                                                               management for distributed
                                                               database environments

                                                     •    Improves performance


    Requires NO changes to your existing infrastructure

5
Pain Points – The Scalability Problem

• Thousands of new online and mobile
  apps launching every day

• Demand climbs for these apps and
  databases can’t keep up

• App must provide uninterrupted
  access and availability

• Database performance and
  scalability is critical



6
Big Data = Big Scaling Needs

       Big Data = Transactions + Interactions + Observations
               Sensors/RFID/Devices      Mobile Web       User Generated Content        Spatial & GPS Coordinates




                                                                                                                            BIG DATA
Petabytes      User Click Stream         Sentiment        Social Interactions & Feeds


               Web Logs               Dynamic Pricing       Search Marketing




                                                                                                 WEB
               Offer History          A/B Testing           Affiliate Networks
Terabytes                                                                                                 External
                                                                                                          Demographics
               Segmentation           Customer Touches




                                                                                 CRM
                                                                                                          Business Data
               Offer Details          Support Contacts                                                    Feeds


Gigabytes
                                                                                                  HD Video, Audio, Images
                                                                                   Behavioral
                                                    ERP


                    Purchase Detail
                                                                                   Targeting      Speech to Text
                    Purchase Record
                                                                                                  Product/Service Logs
                    Payment Record                                                 Dynamic
                                                                                   Funnels
                                                                                                  SMS/MMS
Megabytes



                                      Increasing Data Variety and Complexity

   7
                                           The 451 Group & Teradata
Scalability Pain



Infrastructure
Cost $
                   Large                     You just lost
                   Capital                    customers
                 Expenditure


                                                         Predicted
                                                         Demand

                               Opportunity                   Traditional
                                 Cost                        Hardware

                                                             Actual
                                                             Demand

                                                         Dynamic
                                                         Scaling


                                                                      time


    8
Ongoing “Scaling MySQL” Series

    • August 16 & September 20, 2012
       – Scaling MySQL: ScaleUp versus Scale Out

    • October 23, 2012
       – Methods and challenges to Scale out MySQL

    • December 13, 2012
       – Benefits of Automatic Data Distribution

    • Today
       – Catch 22 of read-write splitting



9
The Database Engine is the Bottleneck...

 • Every write operation is At Least 4 write operations inside the DB:
     – Data segment
     – Index segment
     – Undo segment
     – Transaction log
 • And Multiple Activities in the DB engine memory:
     – Buffer management
     – Locking
     – Thread locks/semaphores
     – Recovery tasks




10
The Database Engine is the Bottleneck

 • Every write operation is At Least 4 write operations inside the DB:
     – Data segment
     – Index segment
     – Undo segment                          Now multiply
    – Transaction log                           by 10TB
                                              accessed by
 • And Multiple Activities in the DB engine memory:
                                                 10000
    – Buffer management
                                              concurrent
    – Locking
                                                sessions
     – Thread locks/semaphores
     – Recovery tasks




11
So… Let’s get to work!

 • I’m growing, have more users, need to support more
   throughput thru scale-out
 • Solution:
     – Create replicated database servers
     – Distribute the sessions across those database servers. Read /
       Write splitting:
          – Reads use the slaves
          – Writes go to the master
 • Benefits:
     –   Better resource consumption
     –   Get more read throughput
     –   Get more write throughput
     –   Awesome!


12
Read / Write Splitting



                          Replication




13
If Done Alone…
• Application code needs to be changed
• Maintain 2 connection pools
      – Write – master
      – Reads – A blend of all slaves
• Every flow, in its beginning, disclaims:
      – “I will do only reads”
      – “I will do reads and writes”
      – What’s the default?
• Result:
      – Writing code, maintaining code
      – Maintaining database ops in the app: add/remove slaves, change
        IPs...
      – Master database is far more occupied than it should be
      – Reads are not well balanced
      – What if replication breaks?
      – Can I read stale data?
 14
Read / Write Splitting with ScaleBase

 •   0 code changes and 0 code maintenance
 •   Reads run faster
 •   Writes run faster
 •   Better resource utilization/load balancing
 •   Improved data consistency/transaction isolation
 •   Built-in failover with high availability
 •   Database aware, replication state aware, replication lag aware
 •   Real time monitoring and alerts
 •   Centralized management dashboard




15
Read / Write Splitting

                             Current




                                              Replication




                          With ScaleBase




                                              Replication




                     Application Experience




16
Read/Write Splitting


                                 Read

                                 Write
             APPLICATION /
     USERS
             WEB SERVERS




                             Replication




17
Scale Out with Amazon AWS RDS Read Replica

                         Current




                                           RDS Read Replicas




                     With ScaleBase




                                           RDS Read Replicas




                  Application Experience




18
Choose Your Scale-out Path


                              Data Distribution


           Database Size



                                      Read/Write Splitting




                           1 DB?
                           Good for me!




                               # of concurrent sessions
19
Scaling Out Achieves Unlimited Scalability

             160000

             140000

             120000

             100000
Throughput




                                                                                               84000
             80000                                                                                     Throughput (TPM)
                                                                                                       Total DB Size (MB)
             60000                                                                60000                # Connections
                                                                     48000
             40000
                                                        36000
                                              24000                                            2500
             20000                                                                2000
                                     12000              1500         1500
                          6000                1000
                 0        500        500
                      1          2           4        6          8           10           14
                                              Number of Databases

     20
Detailed Scale Out Case Studies

     One of world's largest &
     most widely respected
     manufacturers of smart
           phones and
      telecommunications
      hardware & software


                                AppDynamics             Mozilla           Solar Edge
     • Device Apps App          • Next gen APM          • New Product/    • Next Gen
     • Availability               company                 Next Gen App/     Monitoring App
     • Scalability              • Scalability for the     AppStore        • Massive Scale
     • Geo-clustering             Netflix               • Scalability     • Monitors real
                                  implementation        • Geo-sharding      time data from
     • 100 Apps
                                                                            thousands of
     • 300 MySQL DB
                                                                            distributed
                                                                            systems




21
Summary

     • Database scalability is a significant problem
         – App explosion, Big Data, Mobile
     • Scale Up helps somewhat, but Scale Out provides
       a long-term, cost-effective solution

     • ScaleBase has an effective Scale Out
       solution with a proven ROI
         – Improves performance &
           requires NO changes to
           your existing infrastructure
     • Choose your scale-out path....
         – The ScaleBase platform enables
           you to start with R/W splitting and
           grow into automatic data distribution

22
Questions (please enter directly into the GTW side panel)



617.630.2800

www.ScaleBase.com

doron.levari@scalebase.com

paul.campaniello@scalebase.com


23
Thank You
24

Scaling MySQL: Catch 22 of Read Write Splitting

  • 1.
    Webinar: Scaling MySQL Catch22 of Read Write Splitting January 17, 2013
  • 2.
    Agenda 1. Who We Are 2. The Scalability Problem 3. How We Solve it with Read/Write Splitting 4. Customer ROI/Case Studies 5. Q & A (please type questions directly into the GoToWebinar side panel) 2
  • 3.
    Who We Are Presenters: Paul Campaniello, VP of Global Marketing 25 year technology veteran with marketing experience at Mendix, Lumigent, Savantis and Precise. Doron Levari, Founder A technologist and long-time veteran of the database industry. Prior to founding ScaleBase, Doron was CEO to Aluna. 3
  • 4.
    Who We Are ScaleBaseallows apps to cost-effectively scale to an infinite number of users, with NO disruption to the existing infrastructure 4
  • 5.
    The ScaleBase DataTraffic Manager • Database Scalability – Scale out relational databases to unlimited users – Real-time elasticity • Database Availability – Enable high availability of all apps • Centralized Management – Removes complexity and provides a unified point of management for distributed database environments • Improves performance Requires NO changes to your existing infrastructure 5
  • 6.
    Pain Points –The Scalability Problem • Thousands of new online and mobile apps launching every day • Demand climbs for these apps and databases can’t keep up • App must provide uninterrupted access and availability • Database performance and scalability is critical 6
  • 7.
    Big Data =Big Scaling Needs Big Data = Transactions + Interactions + Observations Sensors/RFID/Devices Mobile Web User Generated Content Spatial & GPS Coordinates BIG DATA Petabytes User Click Stream Sentiment Social Interactions & Feeds Web Logs Dynamic Pricing Search Marketing WEB Offer History A/B Testing Affiliate Networks Terabytes External Demographics Segmentation Customer Touches CRM Business Data Offer Details Support Contacts Feeds Gigabytes HD Video, Audio, Images Behavioral ERP Purchase Detail Targeting Speech to Text Purchase Record Product/Service Logs Payment Record Dynamic Funnels SMS/MMS Megabytes Increasing Data Variety and Complexity 7 The 451 Group & Teradata
  • 8.
    Scalability Pain Infrastructure Cost $ Large You just lost Capital customers Expenditure Predicted Demand Opportunity Traditional Cost Hardware Actual Demand Dynamic Scaling time 8
  • 9.
    Ongoing “Scaling MySQL”Series • August 16 & September 20, 2012 – Scaling MySQL: ScaleUp versus Scale Out • October 23, 2012 – Methods and challenges to Scale out MySQL • December 13, 2012 – Benefits of Automatic Data Distribution • Today – Catch 22 of read-write splitting 9
  • 10.
    The Database Engineis the Bottleneck... • Every write operation is At Least 4 write operations inside the DB: – Data segment – Index segment – Undo segment – Transaction log • And Multiple Activities in the DB engine memory: – Buffer management – Locking – Thread locks/semaphores – Recovery tasks 10
  • 11.
    The Database Engineis the Bottleneck • Every write operation is At Least 4 write operations inside the DB: – Data segment – Index segment – Undo segment Now multiply – Transaction log by 10TB accessed by • And Multiple Activities in the DB engine memory: 10000 – Buffer management concurrent – Locking sessions – Thread locks/semaphores – Recovery tasks 11
  • 12.
    So… Let’s getto work! • I’m growing, have more users, need to support more throughput thru scale-out • Solution: – Create replicated database servers – Distribute the sessions across those database servers. Read / Write splitting: – Reads use the slaves – Writes go to the master • Benefits: – Better resource consumption – Get more read throughput – Get more write throughput – Awesome! 12
  • 13.
    Read / WriteSplitting Replication 13
  • 14.
    If Done Alone… •Application code needs to be changed • Maintain 2 connection pools – Write – master – Reads – A blend of all slaves • Every flow, in its beginning, disclaims: – “I will do only reads” – “I will do reads and writes” – What’s the default? • Result: – Writing code, maintaining code – Maintaining database ops in the app: add/remove slaves, change IPs... – Master database is far more occupied than it should be – Reads are not well balanced – What if replication breaks? – Can I read stale data? 14
  • 15.
    Read / WriteSplitting with ScaleBase • 0 code changes and 0 code maintenance • Reads run faster • Writes run faster • Better resource utilization/load balancing • Improved data consistency/transaction isolation • Built-in failover with high availability • Database aware, replication state aware, replication lag aware • Real time monitoring and alerts • Centralized management dashboard 15
  • 16.
    Read / WriteSplitting Current Replication With ScaleBase Replication Application Experience 16
  • 17.
    Read/Write Splitting Read Write APPLICATION / USERS WEB SERVERS Replication 17
  • 18.
    Scale Out withAmazon AWS RDS Read Replica Current RDS Read Replicas With ScaleBase RDS Read Replicas Application Experience 18
  • 19.
    Choose Your Scale-outPath Data Distribution Database Size Read/Write Splitting 1 DB? Good for me! # of concurrent sessions 19
  • 20.
    Scaling Out AchievesUnlimited Scalability 160000 140000 120000 100000 Throughput 84000 80000 Throughput (TPM) Total DB Size (MB) 60000 60000 # Connections 48000 40000 36000 24000 2500 20000 2000 12000 1500 1500 6000 1000 0 500 500 1 2 4 6 8 10 14 Number of Databases 20
  • 21.
    Detailed Scale OutCase Studies One of world's largest & most widely respected manufacturers of smart phones and telecommunications hardware & software AppDynamics Mozilla Solar Edge • Device Apps App • Next gen APM • New Product/ • Next Gen • Availability company Next Gen App/ Monitoring App • Scalability • Scalability for the AppStore • Massive Scale • Geo-clustering Netflix • Scalability • Monitors real implementation • Geo-sharding time data from • 100 Apps thousands of • 300 MySQL DB distributed systems 21
  • 22.
    Summary • Database scalability is a significant problem – App explosion, Big Data, Mobile • Scale Up helps somewhat, but Scale Out provides a long-term, cost-effective solution • ScaleBase has an effective Scale Out solution with a proven ROI – Improves performance & requires NO changes to your existing infrastructure • Choose your scale-out path.... – The ScaleBase platform enables you to start with R/W splitting and grow into automatic data distribution 22
  • 23.
    Questions (please enterdirectly into the GTW side panel) 617.630.2800 www.ScaleBase.com doron.levari@scalebase.com paul.campaniello@scalebase.com 23
  • 24.