1 / 21

Best Practices for Building Tier 1 Enterprise Business Intelligence Solutions with Microsoft SQL Server Analysis Service

DBI407. Best Practices for Building Tier 1 Enterprise Business Intelligence Solutions with Microsoft SQL Server Analysis Services. Adam Jorgensen President Pragmatic Works. What do Tier 1 Solutions have in Common?. Design and Query Considerations High Concurrent User Count

teddy
Download Presentation

Best Practices for Building Tier 1 Enterprise Business Intelligence Solutions with Microsoft SQL Server Analysis Service

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. DBI407 Best Practices for Building Tier 1 Enterprise Business Intelligence Solutions with Microsoft SQL Server Analysis Services Adam Jorgensen President Pragmatic Works

  2. What do Tier 1 Solutions have in Common? • Design and Query Considerations • High Concurrent User Count • Increased Hardware Requirements • IO Hungry! • Real Time Implications • Proven Strategies

  3. Design and Query Considerations • Simplicity is Key • Parent child use • Many to Many • Cell by Cell vs. Block Mode • Data Types • Partitioning and Locking • Threading • Sizing • Distinct Count Performance • Hashing and Locking Pools

  4. Managing Multi User Concurrency • Scaling Out • Synchronization • Read-Only DB’s • SAN Snapshots and Cones • System Engineering • Heap/Virtual Memory

  5. Increasing Hardware Requirements • Using SSD’s • Block Size • Short Stroking • OPAS and Latency • ROLAP (Preview)

  6. Improving IO • Partition Alignment • Storage Enging Performance • ROLAP • Transparent aggregations • No Sub-queries – must use table binding • Disabling the count of rows measure due to subquery • Indexed Views with individual tables • Best Practices Chart

  7. Partitioning Is For Real DW’s

  8. Real Time Implications • Techniques for Real Time • Proactive caching • Rolap • Incremental Processing Molap/Rolap

  9. Proven Strategies – Case Studies • Highlights from SQLCAT (Examples of big installations) • AdCenter • Xbox Live • Yahoo!

  10. Case Study - AdCenter • EMC DMX V-Max to handle the I/O • V-Max are dedicated to the application • Hundreds of disks and spindles devoted to this process • Short-stroking the disks to ensure fastest I/O • Works closely with EMC directly (present at EMC World regularly) • Testing with EMC EFDs (Enterprise Flash Drives) • Systems Engineering team dedicated to DW process • Work closely with various vendors (EMC, HDS, etc.) • References • Accelerating Microsoft adCenter with Microsoft SQL Server 2008 Analysis Services. • REAL PRACTICES: Performance Scaling Microsoft adCenter with Microsoft SQL Server 2008 Analysis Services on EMC VMAX

  11. Ad Center Production

  12. Xbox Live – SSD Performance

  13. Yahoo – Massive Data for Intense Scale Oracle 10g SSAS Cube Builder CDF Partition 1 Partition 1 1.2TB/day Partition 2 Partition 2 12TB cube Partition N Partition N NAS 50GB/hr SSAS Query Servers File1 File2 FileN HW NLB

  14. Summarizing (1/2) • You HAVE to get the design right if you want to scale • Partitioning is absolute crucial • Partition for processing speed • Partition for data latency (real time vs. Stale) • Partition for archival of old data • Secondary concern to above: partition for query speed • Hardware really matters for large cubes • NAND devices are match made in heaven. • Tests show two core Nehalem CPU can often support hundreds of users • Carefully balance IOPS vs. Memory, consider hot portion of cube

  15. Summarizing (2/2) • Locking is your enemy • Consider processing server with sync vs. Individual servers processing • Consider how to flip • For super low latency, ROLAP is the way • Multi User Settings • Just run them as per perf guide • Consider adjusting CoordinatorBalancingFactor • Keep it Simple • Remember the premise: AS is a ”send me any query” database. • You really have to get this right at scale

  16. What’s Next? • Follow, Tweet and Enter to win an Xbox Kinect Bundle • GAME ON! Join us at the top of every hour at the BI booth to compete in the Crescent Puzzle Challenge and Win Prizes • Sign up to be notified when the next CTP is available at: microsoft.com/sqlserver Join the Conversation @MicrosoftBI /MicrosoftBI

  17. Resources • Connect. Share. Discuss. http://northamerica.msteched.com Learning • Sessions On-Demand & Community • Microsoft Certification & Training Resources www.microsoft.com/teched www.microsoft.com/learning • Resources for IT Professionals • Resources for Developers http://microsoft.com/technet http://microsoft.com/msdn

  18. Complete an evaluation on CommNet and enter to win!

More Related