SQL Server 2008 כפלטפורמה לניהול נתונים בעולם ה-Peta byte
What is a VLDB?Used to be a set size:>100mb in the 1980’s
>100gb in the early 1990’s
>1tb in the early 2000’s
>10tb in the mid 2000’s
Now we are dealing with hundreds of TeraBytesNow, a more flexible definition for VLDB prevails:It’s a VLDB when you can no longer use “common and standard” techniques for management and performanceVLDB project ChallengesPlenty of data to handleMassive data insertionMassive query loadsNeed for High AvailabilityNeed for ScalabilityCost reduction
Design Philosophy VLDBDon’t think of it as Large complex database, think of it as smaller manageable componentsPutting all the data in one single database doesn’t make you a hero.Partitioning data into smaller manageable sizesMethods for data partitioning:Server Partitioning
Instance Partitioning
Database Partitioning

Vldb Yaron Moshe

  • 1.
    SQL Server 2008כפלטפורמה לניהול נתונים בעולם ה-Peta byte
  • 2.
    What is aVLDB?Used to be a set size:>100mb in the 1980’s
  • 3.
    >100gb in theearly 1990’s
  • 4.
    >1tb in theearly 2000’s
  • 5.
    >10tb in themid 2000’s
  • 6.
    Now we aredealing with hundreds of TeraBytesNow, a more flexible definition for VLDB prevails:It’s a VLDB when you can no longer use “common and standard” techniques for management and performanceVLDB project ChallengesPlenty of data to handleMassive data insertionMassive query loadsNeed for High AvailabilityNeed for ScalabilityCost reduction
  • 7.
    Design Philosophy VLDBDon’tthink of it as Large complex database, think of it as smaller manageable componentsPutting all the data in one single database doesn’t make you a hero.Partitioning data into smaller manageable sizesMethods for data partitioning:Server Partitioning
  • 8.
  • 9.
  • 10.
    Tables PartitioningData separation:Data Partitioning vs. Function PartitioningExamples of data partitioning by dataData broken up by months in a year: Jan, Feb, Mar……Data broken up by States in the US: Alabama, Arkansas, Alaska…..Examples of data partitioning by functionData separated by time: sales data, consulting data, competitive data….Data separated by business function: manufacturing data, design data, inventory data…….
  • 11.
    Table Partitioning: CommonScenario “Sliding Window”New data is added continually (every minute, hourly, Daily, Weekly, Monthly)Data must be scrubbed, indexed, etc.As data ages, data is modified less frequently and then - not at all.Eventually, it may be discarded Sliding WindowTimeObsoleteHotRead-MostlyRead-Only
  • 12.
    Data Feeder AggrFirsttier Databases AggrTiered ManagerETLQueriesTiered ManagerManager ServerWhere can I find my Data?QueriesSecond tier Databases AggrApplication Server
  • 13.
    FileGroup 2FileGroup 1FileGroup1Key (partitioning column)LocatingARowPartition SchemePartitionFunction
  • 14.
    HA and scalabilityN+1ClusteringSmart management enables fast scalability by adding servers, databases, instances, partitions etc.
  • 15.
    Cost reductionPage LevelData compression – saves up to 50% of disk storage (and reduces IO)Smaller partitions may reduce indexes size by 30% (!) – critical in huge environmentsMoving historical to cheaper storage/media No need for Mega servers - Standard servers are doing the job
  • 16.
    Valinor’s people inthis projectShahar BarMeir DudaiTzahi Hakikat…. And more….
  • 17.