NoSQL and SQL - Why Choose? Enjoy the best of both worlds with MySQL

Copyright © 2014, Oracle and/or its affiliates. All rights reserved. |

NoSQL and SQL: The Best of Both Worlds

Andrew Morgan @andrewmorgan clusterdb.com 18th November 2014


Safe Harbor Statement The following is intended to outline our general product direcPon. It is intended for informaPon purposes only, and may not be incorporated into any contract. It is not a commitment to deliver any material, code, or funcPonality, and should not be relied upon in making purchasing decisions. The development, release, and Pming of any features or funcPonality described for Oracle’s products remains at the sole discrePon of Oracle.



Session Agenda • NoSQL – What are people looking for? • RDBMS – What advantages do they sPll have? • How MySQL Delivers the Best of Both Worlds

– MySQL Cluster •  NoSQL aXributes: Scale-‐out, performance, ease-‐of-‐use, schema flexibility, on-‐line operaPons •  NoSQL APIs

– Key-‐Value store access to InnoDB (Memcached)


Types of NoSQL stores

Key-‐Value

• Cassandra • Memcached • BigTable • Hadoop • Voldermort

Document

• MongoDB • CouchDB

Graph

• Neo4J • FlockDB


Key-‐Value Store



Graph Database


• Massive scalability – No applicaPon-‐level sharding

•  Performance • High Availability/Fault Tolerance •  Ease of use

– Simple operaPons/administraPon – Simple APIs – Quickly evolve applicaPon & schema

Scalability

Performance

HA

Ease of use

What NoSQL must deliver



NoSQL

Simple access paXerns

Compromise on consistency for performance

Ad-‐hoc data format

Simple operaPon

SQL

Complex queries with joins

ACID transacPons

Well defined schemas

Rich set of tools

• No best single soluPon fits all • Mix and match

SPll a role for SQL (RDBMS)?

Scalability

Performance

HA

Ease of use

SQL/Joins

ACID Transac>ons


The TemptaPon to Jejson RelaPonal Model

•  RelaPonal Model: – Data stored in mulPple tables – Many lookups to retrieve a user’s Pmeline

– Is it worth the effort in sejng up this complex data model?

•  Simpler just to store as one document?

The allure of document stores

Examples borrowed from @sarahmei hXps://speakerdeck.com/sarahmei/switching-‐data-‐stores-‐a-‐postmodern-‐comedy

user

friend

post

comment

like

liker

commenter

many

many

many

one

one

many



• Document Model: – EnPre Pmeline in a single document (row)

– Single lookup to retrieve the user’s Pmeline

– Brilliantly efficient model when the document truly contains self-‐contained informaPon

•  Like a real-‐world document!

The allure of document stores {name: ‘Joe’, url: ‘http://clusterdb.com’, stream:[ {friend:{ name: ‘Jane’, posts:[{content: ‘…’, comments:[ {comment: ‘…’, commenter: ‘Joe’}, {…} ], likes: [‘Joe’, ‘Fred’, ‘Billy’] }, {…},{…},… ] }, {…},{…},…}




•  These are all people who have their own data that users will want to view or click through: – Name –  url –  Timeline

•  Easy to represent with FKs in a relaPonal model

But when the data isn’t self contained…

user

friend

post

comment

like

liker

commenter

many

many

many

one

one

many



•  Do I store all data for all friends and likers again at every point they appear in the document?

•  Massive amount of repeated data – Wasted space –  Needs to be kept in sync (and you don’t have transacPons to provide consistency)

The allure of document stores {name: ‘Joe’, url: ‘http://clusterdb.com’, stream:[ {friend:{ name: ‘Jane’, posts:[{content: ‘…’, comments:[ {comment: ‘…’, commenter: ‘Joe’}, {…} ], likes: [‘Joe’, ‘Fred’, ‘Billy’] }, {…},{…},… ] }, {…},{…},…}




•  The reality is that the developer will store the user-‐ids instead

•  Developer is responsible for implemenPng ‘joins’ in the applicaPon –  e.g. scanning all Pmelines for where a user made a comment or liked a post

•  No transacPons to ensure consistency between documents

The allure of document stores {name: 83746251, url: ‘…’ stream:[ {friend:{ name: 9384726153, posts:[{content: ‘…’, comments:[ {comment: ‘…’, commenter: 83746251}, {…} ], likes: [83746251, 750730283, 2938493820] }, {…},{…},… ] }, {…},{…},…}



•  Each aXribute’s labels are stored in every document – More wasted space –  Developers compensate by replacing with short names

The allure of document stores {n5: 83746251, u2: ‘…’ s1:[ {f2:{ n5: 9384726153, p4:[{c4: ‘…’, c3:[ {c2: ‘…’, c8: 83746251}, {…} ], l5: [83746251, 750730283, 2938493820] }, {…},{…},… ] }, {…},{…},…}



MySQL Cluster Overview •  Auto-‐Sharding, MulP-‐Master •  ACID Compliant, OLTP + Real-‐Time AnalyPcs

HIGH SCALE, READS + WRITES

•  Shared nothing, no Single Point of Failure •  Self Healing + On-‐Line OperaPons 99.999% AVAILABILITY

•  In-‐Memory OpPmizaPon + Disk-‐Data •  Predictable Low-‐Latency, Bounded Access Time REAL-‐TIME

•  Key/Value + Complex, RelaPonal Queries •  SQL + Memcached + JavaScript + Java + HTTP/REST & C++ SQL + NoSQL

•  Open Source + Commercial EdiPons •  Commodity hardware + Management, Monitoring Tools LOW TCO


Who’s Using MySQL Cluster?


MySQL Cluster Architecture

MySQL Cluster Data Nodes

Clients

ApplicaPon Layer

Data Layer


MySQL Cluster Scaling


Clients

ApplicaPon Layer

Data Layer



MySQL Cluster HA


Clients

ApplicaPon Layer

Data Layer


On-‐line OperaPons •  Scale the cluster (add & remove nodes on-‐line) • ReparPPon tables • Upgrade / patch servers & OS • Upgrade / patch MySQL Cluster • Back-‐Up •  Evolve the schema on-‐line, in real-‐Pme




1.2 Billion UPDATEs per Minute

• NoSQL C++ API, flexaSynch benchmark

•  30 x Intel E5-‐2600 Intel Servers, 2 socket, 64GB

• ACID TransacPons, with Synchronous ReplicaPon

0

5

10

15

20

25

2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

Mill

ions

of U

PDAT

Es p

er S

econ

d



Join Performance

• Web-‐Based Content Management System – JOINs 11-‐tables, 33.5k rows – Returns 2k rows, 19 columns per row

0

10

20

30

40

50

60

70

80

90

100

MySQL Cluster 7.1 MySQL Cluster 7.2

Query Execution Time Seconds

87.23 seconds

1.26 seconds

70x Faster

Scalability !"

Performance !"

HA !"

Ease of use

SQL/Joins !"

ACID Transac>ons !"


Better performance and operational simplicity MySQL Cluster 7.4.1 DMR

• Performance gain over 7.3 – 47% (Read-‐Only) – 38% (Read-‐Write)

0"

1000"

2000"

3000"

4000"

5000"

6000"

7000"

8000"

9000"

10000"

64" 128" 256" 512"

Tran

sac'on

)per)se

cond

)

Threads)

Sysbench)RO)

7.4.0"

7.3.5"

7.2.16"

0"

1000"

2000"

3000"

4000"

5000"

6000"

7000"

8000"

9000"

64" 128" 256" 512"

Tran

sac'on

s)per)se

cond

)

Threads)

Sysbench)R/W)

7.4.0"

7.3.5"

7.2.16"

•  Faster node restarts



CreaPng & running your first Cluster

Download & Extract • edelivery.oracle.com • www.mysql.com • dev.mysql.com

Configure • Cluster-‐wide “config.ini”

• Per-‐mysqld “my.cnf”

Start processes • Management Nodes • Data Nodes • MySQL Servers


Scalability !"

Performance !"

HA !"

Ease of use !"

SQL/Joins !"

ACID Transac>ons !"



MySQL Cluster Auto-‐Installer

• Fast configuraPon • Auto-‐discovery • Workload opPmized • Repeatable best pracPces

Specify Workload

Auto-‐Discover

Define Topology Deploy



MySQL Cluster Manager 1.  Download MCM/Cluster package from edelivery.oracle.com: 2.  Unzip 3.  Run agent, define, create & start Cluster! $> bin\mcmd --bootstrap MySQL Cluster Manager 1.1.2 started

Connect to MySQL Cluster Manager by running "D:\Andrew\Documents\MySQL\mcm\bin\mcm" -a NOVA:1862

Configuring default cluster 'mycluster'...

Starting default cluster 'mycluster'...

Cluster 'mycluster' started successfully

ndb_mgmd NOVA:1186

ndbd NOVA

ndbd NOVA

mysqld NOVA:3306

mysqld NOVA:3307

ndbapi *

Connect to the database by running "D:\Andrew\Documents\MySQL\mcm\cluster\bin\mysql" -h NOVA -P 3306 -u root


MCM: Upgrade Cluster

mysql> upgrade cluster --package=7.3 mycluster;


Scalability !"

Performance !"

HA !"

Ease of use !"

SQL/Joins !"

ACID Transac>ons !"




NoSQL Access to MySQL Cluster data

Apps Apps Apps Apps Apps Apps Apps Apps Apps Apps Apps Apps

JPA

Cluster JPA

PHP Perl Python Ruby JDBC Cluster J JS Apache Memcached

MySQL JNI Node.JS mod_ndb ndb_eng

NDB API (C++)



Schema-‐Free apps

• Rapid applicaPon evoluPon – New types of data constantly added

– No Pme to get schema extended – Missing skills to extend schema – IniPally roll out to just a few users – Constantly adding to live system


Cluster & Memcached – Schema-‐Free

<town:maidenhead,SL6>

key value


key value

Key Value

town:maidenhead SL6

generic table

ApplicaPon view

SQL view


Cluster & Memcached -‐ Configured Schema


prefix key value


key value

Prefix Table Key-‐col Val-‐col policy

town: map.zip town code cluster

Config tables

town ... code ...

maidenhead ... SL6 ...

map.zip

ApplicaPon view

SQL view


MySQL 5.6 Memcached with InnoDB

0

10000

20000

30000

40000

50000

60000

70000

80000

8 32 128 512

TPS

Client Connections

Memcached API

SQL

Clients and Applica>ons

MySQL Server Memcached Plug-‐in

innodb_ memcached

local cache (opPonal)

Handler API InnoDB API

InnoDB Storage Engine

mysqld process

SQL Memcached Protocol

Up to 9x Higher “SET / INSERT” Throughput


Node.js NoSQL API • NaPve JavaScript access to MySQL Cluster – End-‐to-‐End JavaScript: browser to the app & DB

– Storing and retrieving JavaScript objects directly in MySQL Cluster

– Eliminate SQL transformaPon

•  Implemented as a module for node.js – Integrates Cluster API library within the web app

•  Couple high performance, distributed apps, with high performance distributed database

• OpPonally routes through MySQL Server – Use with InnoDB

V8 JavaScript Engine

MySQL Cluster Node.js Module


Clients


NoSQL API for Node.js & FKs

FKs enforced on all APIs: { message: 'Error', sqlstate: '23000', ndb_error: null, cause: {message: 'Foreign key constraint violated: No parent row found [255]', sqlstate: '23000', ndb_error: { message: 'Foreign key constraint violated: No parent row found', code: 255, classification: 'ConstraintViolation', handler_error_code: 151, status: 'PermanentError' }, cause: null } }


Oracle University MySQL Training Services Prepare Your Organization to Enable Reliable and High-Performance Web-Based Database Applications

“Training and team skill have the most significant impact on overall performance of technology and success of technology projects.” - IDC, 2013

Premier Support customers eligible to save 20% on learning credits.

Benefits §  Expert-led training to support your MySQL learning needs §  Flexibility to train in the classroom or online §  Hands-on experience to gain real world experience §  Key skills needed for database administrators and developers

• MySQL for Beginners § MySQL for Database Administrators § MySQL Performance Tuning § MySQL Cluster – NEW - Register Your Interest! § MySQL and PHP - Developing Dynamic Web Applications § MySQL for Developers § MySQL Developer Techniques § MySQL 5.6 Database Administrator § MySQL 5.6 Developer To find out more about available MySQL Training & Certification

offerings, go to: education.oracle.com/mysql

Top Courses for Administrators and Developers

Top Certifications

RECENTLY RELEASED !!ALL NEW!! MySQL Cluster Training To Register your interest to influence the

schedule on this newly released course – go to educaPon.oracle.com/mysql and click on the

MySQL Cluster Course


• Tomas Ulin – Vice President of the MySQL Engineering Team

• Oracle DBAs, Make the Most of MySQL

• Wednesday 11:00-‐11:45

MySQL Keynote @ DOAG2014


Next Steps Learn More •  www.mysql.com/cluster •  AuthenPc MySQL Curriculum: hXp://oracle.com/educaPon/mysql

Try it Out •  dev.mysql.com/downloads/cluster/

Let us know what you think •  clusterdb.com • @clusterdb •  forums.mysql.com/list.php?25

Software

NoSQL and SQL - Why Choose? Enjoy the best of both worlds with MySQL