What's New:
MySQL Cluster 7.0 Webinar
Matthew Montgomery
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 1
Agenda
Industry Trends MySQL in Communications MySQL Cluster Architecture MySQL Cluster 7.0 Use Cases Resources
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 2
About MySQL
14 Years of Development Acquired by Sun in February 2008 400+ in Database Group 750+ Partners 70K+ Downloads Per Day Customers across every major operating system, hardware vendor, geography, industry, and application type
High Performance Reliable Easy to Use
Copyright 2009 MySQL Sun Microsystems. The Worlds Most Popular Open Source Database 3
Where is MySQL Located? Everywhere!
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 4
MySQL Industry Adoption
MySQL is the third most commonly used RDBMS #1 open-source database
Copyright 2009 MySQL Sun Microsystems. The Worlds Most Popular Open Source Database 5
Serving Key Markets & Industry Leaders
Web / Web 2.0
OEM /ISV's
On Demand, SaaS, Hosting
Telecommunications
Enterprise 2.0
Open-source powers the Web & the Network
Copyright 2009 MySQL Sun Microsystems. The Worlds Most Popular Open Source Database 6
MySQL in Communications
http://www.mysql.com/industry/communications/resources.html#customer_case_studies
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 7
Key Opportunity: Service Convergence
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 8
MySQL Cluster Customers and Applications
NEPs, ISVs and Communications Service Providers Network Subscriber databases Service Delivery Platforms Telco Application Servers VoIP Infrastructure Intelligent Networks Value-added services IPTV / VoD Mobile Content Delivery On-Line app stores and portals DNS/DHCP for Broadband Payment Gateways Data Store for LDAP Directories
http://www.mysql.com/industry/telecom/
Copyright 2009 MySQL Sun Microsystems. The Worlds Most Popular Open Source Database 9
> Messaging / Presence
MySQL Cluster
Key Capabilities
Carrier-grade database Highly available
Shared-nothing, ACID-compliant relational database Five 9s (99.999%) availability using commodity hardware. Self-healing, sub-second failover High-load, real-time performance Predictable low latency, bounded access times Incrementally scale up, out and on-line as application demands Linearly scale with distribution awareness Open source, multiple data access High-performance APIs (C++/Java), SQL, LDAP, Web Services
The Worlds Most Popular Open Source Database 10
Real-time performance Linearly & dynamically scalable Open development
Copyright 2009 MySQL Sun Microsystems.
MySQL Cluster Architecture
Clients
MySQL Cluster Application Nodes
LDAP
MGM
MGM
MySQL Cluster Data (NDBD) Nodes
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 11
MySQL Cluster Geographic Redundancy
Asynchronous replication
NDB API
NDB API
NDB storage engine
NDB storage engine
Cluster 1
Cluster 2
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 12
MySQL Cluster Carrier Grade Edition 7
Key Enhancements
Enhanced Performance & Scalability
Multi-Threaded Data
Expanded Platforms & Interoperability
Windows Port Data Store for LDAP
Simplified Maintenance
Back-up Snapshot
Nodes
Dynamic On-Line
Directories
Scalability
Large Record
Handling
Multi-Threaded Disk
Data File Access
Announced at MySQL UC, April 21st 2009
Copyright 2009 MySQL Sun Microsystems. The Worlds Most Popular Open Source Database 13
"We view MySQL Cluster Carrier Grade Edition as a strategic technology in our applications portfolio. With MySQL Cluster CGE 7.0 we are enabling our customers to expose the rich capabilities of their networks to drive a long-tailed innovation model. As a result, the operator can leverage the creativity and agility of the web ecosystem to deliver new compelling, personalized and targeted services to their subscribers without compromising reliability and scalability."
Jeff Cortley, VP Subscriber Data Management, Alcatel-Lucent
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 14
Online Add Node (1) Add Node Group
Authid (PK) 1 2 3 4 Frame Albert Ernest Johann Junichiro Iname Camus Hemingway Goethe Tanizaki Country France USA Germany Japan
Application
Node group
New node group
Authid (PK) 1 2 3 4
Frame Albert Ernest Johann Junichiro
Iname Camus Hemingway Goethe Tanizaki
Country France USA Germany Japan
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 15
Online Add Node (2) Copy Data
Authid (PK) 1 2 3 4 Frame Albert Ernest Johann Junichiro Iname Camus Hemingway Goethe Tanizaki Country France USA Germany Japan
Application
Node Group
New Node Group
Authid (PK) 1 2 3 4
Frame Albert Ernest Johann Junichiro
Iname Camus Hemingway Goethe Tanizaki
Country France USA Germany Japan Authid (PK) 2 4 Frame Ernest Junichiro Iname Hemingway Tanizaki Country USA Japan
No extra space needed on existing nodes!
The Worlds Most Popular Open Source Database 16
Copyright 2009 MySQL Sun Microsystems.
Online Add Node (3) Switch Distribution
Authid (PK) 1 2 3 4 Frame Albert Ernest Johann Junichiro Iname Camus Hemingway Goethe Tanizaki Country France USA Germany Japan
Application
Node group 1
Node group 2
Authid (PK) 1 2 3 4
Frame Albert Ernest Johann Junichiro
Iname Camus Hemingway Goethe Tanizaki
Country France USA Germany Japan Authid (PK) 2 4 Frame Ernest Junichiro Iname Hemingway Tanizaki Country USA Japan
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 17
Online Add Node (4) Delete Rows
Dynamic scaling of a running cluster with no interruption to service
Application
Authid (PK) 1 2 3 4 Frame Albert Ernest Johann Junichiro Iname Camus Hemingway Goethe Tanizaki Country France USA Germany Japan
Node group 1
Node group 2
Authid (PK) 1 3
Frame Albert Johann
Iname Camus Goethe
Country France Germany
Authid (PK) 2 4
Frame Ernest Junichiro
Iname Hemingway Tanizaki
Country USA Japan
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 18
Scale Up: Multi-Threaded Data Nodes
TC LQH LQH LQH LQH
ACC ACC ACC ACC Index Index Index Index memor memory memor memor yy y TUP TUP TUP TUP Data Data Data Data memor memory memor memor yy y
Multi threaded ndbd Cluster 7.0 Sub partition threads (up to 4) Separate TC and communication threads Utilizes 8 cores
Improved Vertical Scalability on Multi-Core / Multi-Thread Hardware
- Increase Cluster throughput - Reduce hardware requirements to achieve equivalent throughput
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 19
MySQL Cluster Benchmarks
DBT2 Benchmark, 4-MySQL Cluster Data Nodes
300000
Transactions Per Minute
250000 200000 150000 100000 50000 0 1 4 8 12 16 20 24 28 32 36 40
MySQL Cluster 7.0 MySQL Cluster 6.3
Data Nodes
Sun Fire x4450s
SQL Nodes
Sun Fire x4600s & x4450s
Number of MySQL Server Nodes
MySQL Cluster delivers
OpenSolaris Gigabit Ethernet
4.3x higher throughput at 4-nodes MySQL Cluster 7 achieves 140k+ TPM with 2 data nodes vs 8 data nodes required for MySQL Cluster 6.3 4x less power and space consumption
Higher cluster throughput with fewer nodes
Copyright 2009 MySQL Sun Microsystems. The Worlds Most Popular Open Source Database 20
Multithreaded Disk Data Access
Operation s requiring I/O
All file access is through I/O threads Previously a one-to-one mapping between all open files and I/O threads > Still used for some files where latency of single I/O critical e.g., redo logs Can now use pooled I/O threads, breaking the one-to-one mapping > Greater I/O throughput to the file > Results in increased performance for disk-based table data > Users no longer need to split large files into smaller ones as a workaround saving admin work > Extends to other files to reduce overall number of I/O threads in the system saving memory
I/O threads
Files
Single-threaded disk access
Multithreaded disk access
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 21
Large Record Handling (1)
App
4 bytes
Col id 1 Col id 2 Col id 3
NDB
4 bytes
1 2 3
App
bitmap
NDB
1 2 3
Optimization of the NDB messaging protocols between apps and NDB nodes and between NDB nodes Packed read reduces the size of read and response messages > Reduces bandwidth used on the network increasing throughput Read request: replaces a list of column identifiers with a bitmap Read response: column values no longer need to be packed to a 4 byte boundary Changes are transparent to the application, but designers might choose to design schemas that further exploit the mechanism so that fields pack tightly
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 22
Large Record Handling (2)
Ap p
TCKEYRE Q KEYINF O KEYINF O ATTRINF O ATTRINF O
<= 100 bytes
NDB
Ap p
TCKEYRE Q KEYINFO ATTRINFO
NDB
Optimization of the NDB messaging protocols between apps and NDB nodes and between NDB nodes Long signal transactions: >Reduces bandwidth used on the network, increasing throughput (50100% improvements observed) Transparent to the application
Copyright 2009 MySQL Sun Microsystems.
Existing messages limited to 100 bytes so complex requests split across multiple messages: > Bandwidth wasted on protocol overhead (header info) for each message > CPU consumed reassembling operation LST enables large operations to be specified in a single message (up to 32 KB); fragmented signals can be used if larger messages are needed
The Worlds Most Popular Open Source Database 23
MySQL Cluster as a Directory Data Store
Designed for large, complex directories with high update rates Preserves and enhances investment in existing LDAP Directories
Enables user to take advantage of MySQL Cluster to provide high availability, massively scaleable, real time clustered database, using driver from directory vendor Maintain standard API data access of LDAP Maintain directory benefits while improving agility, Copyright 2009 reducing database management complexity & Open Source Database 24 MySQL Sun Microsystems. The Worlds Most Popular cost
MySQL Cluster CGE 7.0 Key Features
E n O L I S M P E n h a n c e d P e r f o r m d a e n c e & S c a e l a l y b e i l it y x p : l o i t i n g d b ms i nt d g l e d a t a n o - l ni n e - n a o d d d e w i t h r e p a r g e r e c o r d h a n d m p r o v e m - n e o n d t as l t o o i c p e e d s u p i n s e r t s u- T l th i r e a d e d D is k D a r a l l e l a c c e s s t o x p a n d e d P la t f o r m e f f e a r t it io n l i n g nm t em r u n / u p d a t a t a A c t h e- b sa as s & I n c t i v i n g i c a e s c e m de e r o
t i o n s / r e a d s o f l a r g e s s t fa i l b e l se ; d s ap t e a e a d c s c p e r a b i li t y
t e
W i n d o w s - Gs u A p ) p o r t ( p r e S u p p o r t f o r C - lG u sr a t e d r e a D s i a- E e Cnc da o r r r y i e B r a c k r t R e - t ia m l e c lu s t e r e d d a t a b a s e u s i n g d r i v S im p li f i e h o a o d M o n i t o r i n g & M a n a g e t a m c e a n c t h
f r o
S n a p s N D B C s c h e m
t o p t i o n f o r b a c k u p s n s i s t e c no tn c f o g n u f ri ga u i or a n t i do na i t c h a n g e s
&
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 25
Management System Improved Management
Centralized deployment Upgrade
> Drop in new version
Today, need to write application logic for replication, such as:
> synchronizing a slave > detection link failure > link failover > master promotion > others
Cluster reconfiguration
> Rolling restart
Cluster monitoring
> Recovery actions
Multi-site management
MySQL Cluster Manager 1.0 GA second half CY09 will deliver a subset of this functionality
Simplified replication management with automation and optimization of:
> Slave synchronization > Replication monitoring > Link failover > Resynchronization after site
failover
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 26
Alcatel-Lucent: Real-Time Platform
Applications
Infrastructure for next-generation converged services including subscriber data management (HLR, HSS) and the XDMS contacts and presence management SDP
Key business benefits
Only MySQL Cluster CGE offers the scalability and the high availability required for the system
Why MySQL?
http://mysql.com/customers/view/? id=566
Performance Reliability Lower costs
MySQL Cluster won the performance test hands-down, and it fitted our needs perfectly. We evaluated shared-disk clustered databases, but the cost would
have been at least 10x more.
Copyright 2009 MySQL Sun Microsystems.
Franois Leygues, Systems Manager, Alcatel-Lucent Source Database 27 The Worlds Most Popular Open
27
UTStarcom
Applications
IPTV with data base requirements for eCommerce, OLTP, and data warehousing
Key Business Benefits
Unlock new revenue streams for operators with bundled IPTV offering Delivered high quality solution to millions of subscribers
Why MySQL?
Ensured 99.999% availability High performance and scalability
As one of the leading databases, MySQL Cluster provides high availability, reliable services, and proven ability to scale, satisfying our telco requirements.
Copyright 2009 MySQL Sun Microsystems.
Tommy L, Senior Product Manager, UTStarcom
28
The Worlds Most Popular Open Source Database 28
M1 Singapore
Application
> RewardsPlus enables M1 to run credit expiration campaigns, generate comparison reports on promotional efficiency and send customized SMS messages to subscribers.
Key business benefit
> M1 has improved customer loyalty in a highly competitive market where delivering new value-added services and growing your subscriber base are key to success.
Why MySQL?
> Reliability & Scale Out
We rely on MySQL Cluster as the key technology behind RewardsPlus, which helps us increase the average revenue per user, and improve customer loyalty.
Tan Tiong Heng, General Manager, Core Network Planning, M1, Singapore
Copyright 2009 MySQL Sun Microsystems.
29
The Worlds Most Popular Open Source Database 29
Resources to Get Started
Getting Started with MySQL Cluster
http://www.mysql.com/products/database/cluster/get-started.html
Download MySQL Cluster
http://dev.mysql.com/downloads/cluster/7.0.html
MySQL Cluster on the Web
http://www.mysql.com/products/database/cluster/
MySQL Cluster 7.0, Architecture and New Features
http://www.mysql.com/why-mysql/whitepapers/mysql_wp_cluster7_architecture.php
MySQL Cluster 7.0 Evaluation Guide
http://www.mysql.com/why-mysql/whitepapers/mysql_cluster_eval_guide.php
MySQL in Communications
http://www.mysql.com/industry/communications/
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 30
Questions?
Matthew Montgomery
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 31
Copyright 2009 MySQL Sun Microsystems.
The Worlds Most Popular Open Source Database 32
