MYSQL

Loading...

Flash Player 9 (or above) is needed to view presentations.
We have detected that you do not have it on your computer. To install it, go here.

0 comments

Post a comment

    Post a comment
    Embed Video
    Edit your comment Cancel

    4 Favorites

    MYSQL - Presentation Transcript

    1. MySQL Cluster Carrier Grade Edition
    2. MySQL Cluster Carrier Grade Edition
      • Provides the telecom industry with a true real-time database with the flexibility of a relational database plus the cost savings of open source
      • Designed by telecom for telecom
      • Ideally suited for the telecom industry's top carriers, operators and equipment manufactures
      • Used by industry leaders for real-time, scalable data management
    3. Basic Cluster Architecture Application Nodes (NDB API and/or MySQL Server) ‏ Data Nodes Management Nodes Clients NDB API NDB Storage Engine
    4. Geographic Redundancy For more information on MySQL Cluster and Replication visit: http://dev.mysql.com/doc/refman/5.1/en/mysql-cluster-replication.html
      • Asynchronous MySQL Replication between Clusters
      • Enables a higher degree of availability and continuity
      Replication Cluster 1 Cluster 2
    5. Carrier Grade Edition Overview Geographic Replication Scalability Performance High Availability Requirement
      • Increase availability/scalability across geographies & data centers
      • Mitigate potential network or site failures
      • Enables Scale-Out
      • Asynchronous data replication
      • Incrementally “scale-out” for increased capacity, performance & scalability
      • Linear scalability with user-defined partitioning
      • Support for disk-based data and large objects
      • In-memory and disk-based data storage
      • Fast, native data access via NDB API
      • Distributed, synchronous, “shared-nothing” data storage
      • Automated fail-over in under a second
      • Automatic data resynchronization in the event of a failure
      • Online backups
      • Rolling upgrades
      Feature Set
    6. Carrier Grade Edition Overview Lower TCO Rapid Development Interoperability Reliability Requirement
      • Runs on commodity, Advanced TCA hardware
      • Lower cost licensing model
      • Minimized design complexity and associated maintenance costs
      • Based on open source and open standards
      • SQL, Java and NDB API data access
      • Automatic and user-defined partitioning
      • Professional consulting and customized application development from MySQL
      • Leverage telecom specific features
      • Most popular open source database in the world
      • Runs on popular open source and proprietary technology stacks
      • Support for ACID transactions
      • Synchronous intra-cluster data replication
      • Local and Global checkpoints
      • 24x7 Technical Support
      • Professional training and certification
      Feature Set
    7. What’s Different? Adaptability Performance Scalability Quality & Reliability Stability Characteristic
      • Online Add Column
      • Support for mixed (x86, SPARC, PowerPC) environments
      • Optimizations to communication protocol enabling API to Data Node over WAN
      • More efficient Client/Server data handling and transfers
      • Other optimizations (Memory, BLOBs, Disk-based data, etc.) ‏
      • Support for larger clusters, up to 255 Nodes
      • Conflict detection support in replication
      • Circular replication support
      • Support for user-defined partitioning
      • Developed & tested with major TEMs
      • Separate development tree from MySQL Server
      • Permits prioritized bug fixes & enhancements for customers
      Benefits
    8. System Requirements COTS – Advanced TCA 32 & 64-bit x86, PowerPC & SPARC Hardware Varies on size of database, # of hosts, # of replicas Memory Linux (Red Hat, SuSE), Solaris, HP-UX, Mac OSX, others… Operating System 100 Mbps or better (Gigabit recommended, SCI supported) ‏ Network Shared-Nothing - Memory & Disk Data SCSI or RAID for I/O performance Storage Maximum of 255 total nodes (48 Data Nodes) ‏ Hosts Requirement System Component
    9. NDB Storage Engine
      • Memory and disk-based
      • ACID Transactions
      • Row-Level Locks
      • Synchronous replication across Data Nodes
        • Linear Two Phase Commit
      • Data automatically or user partitioned
      • Local & Global Checkpoints to disk for persistence
      • Automatic fail over and data resynchronization
      • Data can be accessed via SQL or NDB API
      • Online Backups
      Currently not supported: - Foreign Keys - Full-text Search - Microsoft Windows
    10. Data Distribution & Partitioning P1 - Primary P2 - Replica P1 - Replica P2 - Primary data data data data col2 data data 4 data data 3 data data 2 data data 1 col3 col1 PK Two Data Nodes * Complete copy of database on each node Two Replicas * Two copies of each partition One Node Group * One set of identical Data Nodes Partitions - Automatic or User-Defined - Horizontal - MD5 hash of Primary Key Node Group
    11. Data Distribution & Partitioning P1 - Primary P2 - Replica P1 - Replica P2 - Primary P3 - Primary P4 - Replica P3 - Replica P4 - Primary Four Data Nodes * Half of database on each node Two Replicas * Two copies of each partition Two Node Groups * Two sets of identical Data Nodes Node Group 0 Node Group 1 data data data data col2 data data 4 data data 3 data data 2 data data 1 col3 col1 PK
    12. Design Considerations
      • Number of Replicas
        • Up to 4 supported, 2 is most common
        • NoOfReplicas parameter in config.ini
      • Partitions
        • Automatic
        • Horizontal
        • MD5 hash of primary key
      • Memory Sizing
        • Leverage ndb_size.pl script
        • Data Size * Replicas * 1.1 = Total Memory
        • Example: 2 GB * 2 * 1.1 = 4.4 GB
        • (Data Size * Replicas * 1.1)/Nodes = RAM Per Node
        • Example: (2 GB * 2 * 1.1)/4 = 1.1 GB
      • Disk Storage
        • Used for data, backups, checkpoints and logs
        • SCSI or RAID for faster backups or write-heavy apps
    13. Detecting Lost Connections - Heartbeats
    14. Data Node Failure & Recovery
      • Server 3 failure is detected
      • Server 4 handles Server 3 requests
      P1 - Primary P2 - Replica P1 - Replica P2 - Primary P3 - Primary P4 - Replica P3 - Replica P4 - Primary 1 3 2 4 Node Group 0 Node Group 1
    15. Data Node Failure & Recovery
      • Server 3 begins copying data from Server 4
      • Complete or just changes since failure
      P1 - Primary P2 - Replica P1 - Replica P2 - Primary P3 - Primary P4 - Replica P3 - Replica P4 - Primary 1 3 2 4 Node Group 0 Node Group 1
    16. Data Node Failure & Recovery
      • Data is available as it is restored
      • Server 3 rejoins the Cluster
      P1 - Primary P2 - Replica P1 - Replica P2 - Primary P3 - Primary P4 - Replica P3 - Replica P4 - Primary 1 3 2 4 Node Group 0 Node Group 1
    17. Network Partitioning – Avoiding “Split Brain”
      • By default, the Management Node acts as arbitrator
      • In the event of a network failure, arbitration determines which set of nodes should continue while shutting down the rest
      X Application Nodes (NDB API and/or MySQL Server) ‏ Data Nodes Management Nodes NDB API
    18. Queries and Transactions
      • MySQL Server is a SQL interface for the NDB storage engine
      • NDB API for fast, native data access
      • Data Node answers query or with other Nodes in the Node Group
        • Round-robin by default
        • Transaction Coordinator
        • Local Query Handler
      • With write or updates, a Data Node initiates the transaction on all Nodes in the Node Group
      • Linear two-phase commit
        • Ensures all Data Nodes have received and acted on a “write” before committing
    19. MySQL Cluster NDB API
      • High performance C++ API for NDB storage engine
      • Implements indexes, scans, transactions & events
      • ACID-compliant
      • Object-oriented error-handling
      • Also, MGM API for administrative tasks
      For more information about the NDB API Developers’ Guide visit: http://dev.mysql.com/doc/ndbapi/en/index.html
    20. Linear Scalability
      • Primary Key look-ups are very efficient
        • Increase the number of PK operations in a transaction, batch size
        • Yields linear improvements in performance
      • Multiple-record look-ups benefit from user-defined partitioning
      • Limit full table scans and use of JOINs
    21. Linear Scalability Benefits
    22. Reliability: Checkpoints
      • Process of writing data and logs from memory to disk
      • Local Checkpoint (LCP) ‏
        • Take place for all Data Nodes in the cluster concurrently
        • Persists all of a Data Node's data to disk
        • Interval varies by amount of data and level of cluster activity
        • UNDO Log used to create consistent copy without any locks
        • LCP allows Data Node to remove old REDO Logs
      LCP UNDO Log
    23. Reliability: Checkpoints
      • Global Checkpoint (GCP) ‏
        • Transactions on all nodes are synchronized
        • Recent transactions stored in REDO Log flushed to disk
        • Frequency controlled by TimeBetweenGlobalCheckpoints
        • Performance vs. durability
      GCP REDO Log GCP Transactions Synchronized
    24. Where is MySQL Cluster CGE used
      • Subscriber Databases Management
      • Web Session Management
      • Service Delivery Platforms
        • Messaging
        • Presence
      • VoIP Infrastructure
      • Intelligent Networks
      • Value Added Services
      • IPTV
      • Mobile Content Delivery
      • Gaming
      • DNS/DHCP for Broadband
      • Payment Gateways
      • Directory Services (LDAP)
      http://www.mysql.com/industry/telecom/
      • Nokia Siemens Networks, Alcatel Lucent, Ericsson, Nortel, HP OpenCall, Cisco, Motorola
        • Areas include subscriber repositories, SIP infrastructure, Service Delivery Platforms, Prepaid
      • Vodafone UK, Deutsche Telekom, Orange/FT, Telenor, Telekom Austria
        • Areas include subscriber directories, “modern” IN platforms (SEPs), new services (VoIP, Video/IP), DHCP infrastructure for Broadband Networks, Payment Gateways, Mediation
      Customer Snapshots
    25. Thank You

    + gilashikwagilashikwa, 3 years ago

    custom

    1862 views, 4 favs, 0 embeds more stats

    MySQL Cluster Carrier Grade Edition

    More info about this document

    © All Rights Reserved

    Go to text version

    • Total Views 1862
      • 1862 on SlideShare
      • 0 from embeds
    • Comments 0
    • Favorites 4
    • Downloads 125
    Most viewed embeds

    more

    All embeds

    less

    Flagged as inappropriate Flag as inappropriate
    Flag as inappropriate

    Select your reason for flagging this presentation as inappropriate. If needed, use the feedback form to let us know more details.

    Cancel
    File a copyright complaint
    Having problems? Go to our helpdesk?

    Categories