An introduction to column store indexes and batch mode
Download as PPTX, PDF2 likes1,345 views
This document discusses column store databases and how they work. It explains that column store databases store data by column rather than row to better utilize modern CPU architectures. It describes how column stores use compression techniques like run-length encoding and dictionaries. It also demonstrates how batch processing and sorting data can improve performance of queries against column stores by keeping more data in CPU caches.
Ad
More Related Content
What's hot (20)
Viewers also liked (20)
Ad
Similar to An introduction to column store indexes and batch mode (20)
![JavaOne2016 - Microservices: Terabytes in Microseconds [CON4516]](https://cdn.slidesharecdn.com/ss_thumbnails/javaone2016microservicesfinal-160922163749-thumbnail.jpg?width=560&fit=bounds)
![JavaOne2016 - Microservices: Terabytes in Microseconds [CON4516]](https://cdn.slidesharecdn.com/ss_thumbnails/javaone2016microservicesfinal-160922165618-thumbnail.jpg?width=560&fit=bounds)
Ad
More from Chris Adkin (9)
Recently uploaded (20)
An introduction to column store indexes and batch mode
- 1. + CPU DBA Level 300
- 2. About me An independent SQL Consultant A user of SQL Server from version 2000 onwards with 12+ years experience. I have a passion for understanding how the database engine works at a deep level.
- 3. A Brief History Of Column Store Technology The lineage of column store databases can be traced back to the MonetDb and VectorWise projects from Holland, developed at around the turn of the millennium. Store is column oriented. Column store technology aims to exploit modern CPU architectures. Virtually all database vendors now have a column store database offering. Many people predict a future where all OLAP workloads will be serviced by column oriented databases.
- 4. Column Store Compression Schemes Colour Red Red Blue Blue Green Green Green Dictionary Lookup ID Label 1 Red 2 Blue 3 Green Segment Lookup ID Run Length 1 2 2 2 3 3 Compressing data going down the column using run length compression. Global and local dictionaries are used to store compression metadata.
- 5. Column store segments Column Store Index ‘Anantomy’ Global dictionary Deletion Bitmap Local Dictionary
- 6. What Levels Of Compression Can Be Achieved ? 350 300 250 200 150 100 50 0 Heap Row Compression Page compression Clustered column store index Clustered column store index archive compression * Posts tables from the four largest stack exchanges combined ( superuser, serverfault, maths and Ubuntu ) 53 % 59 % 64 % 72 %
- 7. Demonstration 1: The Difference Batch Mode Makes Test Data Creation
- 8. Demonstration 1: The Difference Batch Mode Makes Test Queries
- 9. How Queries are Executedlans Run Row by row Row by row Row by row Row by row How do rows travel between Iterators ? Control flow Data Flow
- 10. L0 UOP cache Core Modern CPU Architecture C P U L3 Cache L1 Instruction Cache 32KB L2 Unified Cache 256K Core L0 UOP cache L1 Data Cache 32KB Power and Clock Un-core QPI Memory Controller L1 Data Cache 32KB Core Core L1 Instruction Cache 32KB L2 Unified Cache 256K Bi-directional ring bus system-on-chip ( SOC ) design with CPU cores as the basic building block. Utility services are provisioned by the ‘Un-core’ part of the CPU die. Four level cache hierarchy. Memory bus TLB IO … QPI
- 11. Memory Is The “New Disk”Memory 4 4 4 18 14 11 11 11 38 167 0 50 100 150 200 Main memory L3 Cache Full Random access L3 Cache In Page Random access L3 Cache sequential access L2 Cache Full Random access L2 Cache In Page Random access L2 Cache sequential access L1 Cache In Full Random access L1 Cache In Page Random access L1 Cache sequential access Batch mode is about working in the 4 ~ 38 clock cycle range and NOT the 167 cycle “CPU stall” range.
- 12. How Can A Column Store Index Fit Inside The CPU Cache ? C P U Column store object pool Segment Batches
- 13. The Column Store Object Pool
- 14. Batch Mode Pre-Requisites Feature SQL Server 2012 SQL Server 2014 Presence of column store indexes Yes Yes Parallel execution plan Yes Yes No outer joins, NOT Ins or UNION ALLs Yes No Hash joins do not spill from memory Yes No Scalar aggregates cannot be used Yes No
- 15. Feature SQL Server 2012 SQL Server 2014 Column store indexes Yes Yes Clustered column store indexes No Yes Updateable column store indexes No Yes Column store archive compression No Yes Columns in a column store index can be dropped No Yes Support for GUID, binary, datetimeoffset precision > 2, numeric precision > 18. No Yes Enhanced compression by storing short strings natively ( instead of 32 bit IDs ) No Yes Bookmark support ( row_group_id:tuple_id) No Yes Mixed row / batch mode execution No Yes Optimized hash build and join in a single iterator No Yes Hash memory spills cause row mode execution No Yes Iterators supported Scan, filter, project, hash (inner) join and (local) hash aggregate Yes
- 16. How Column Store Index Updates Are Handled Row Groups A B C Columns < 1,048,576 rows Tuple mover Encode and Compress Segments Store Blobs Encode & Compress Delta stores
- 17. aDemonstration 2: Delta Stores In Action
- 18. Demonstration 3: Pre-sorting and Segment Elimination Test Data Creation
- 19. Demonstration 3: Pre-sorting and Segment Elimination Test Queries
- 20. Demonstration 4: Pre-sorting and Hash Aggregate Performance
- 21. Test Setup CPU 6 core 2.0 Ghz (Sandybridge) Warm large object cache used in all tests to remove storage as a factor. CPU 6 core 2.0 Ghz (Sandybridge) 48 Gb quad channel 1333 Mhz DDR3 memory Hyper-threading enabled, unless specified otherwise.
- 22. Atypical Data Warehouse Query On Extra Large Non Sorted Data 1095500000 rows 1,798MB in size
- 23. Atypical Data Warehouse Query On Extra Large Pre-Sorted Data 1095500000 rows 8,555MB in size
- 24. Elapsed Time (ms) / Degree of Parallelism 80000 70000 60000 50000 40000 30000 20000 10000 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (ms) Degree of Parallelism Non-sorted column store Sorted column store
- 25. Lowering Clock Cycles Per Instruction By Leveraging SIMD 1 + 2 = 3 Scalar instruction 1 2 3 4 + 2 3 4 5 = 3 5 7 9 C = A + B SIMD instruction Vector C = Vector A + Vector B
- 26. Takeaways Column store indexes are only half the story, its column store index and batch mode that make the real difference to performance. Pre-sort data where applicable and possible to encourage segment elimination. Pre-sort data on fact table key column subject to the heaviest hash join / aggregate activity. Column Store indexes and batch mode is fast, but not scalable. Many other vendors leverage SIMD, Microsoft are yet to do this, this can result in another step change in performance.
- 27. Questions ?
Editor's Notes
- #5: Run length compression is a generic term that pre-dates the column store functionality in SQL Server, it alludes to the techniques of compressing data by converting sequences of values into encoded “Run lengths”. The database engine scans down a column and stores each unique value it encounters in a dictionary, this can be local to a segment, ( the basic column store unit of storage; containing roughly 1 million rows ) and / or the dictionary can be global to the column store. Where sequences of values are found these are stored as encoded run lengths. In the example above the sequence of two red values is stored as 1, 2 etc . . .
- #10: Each operator has an open(), close(), next() method. An operator pulls data through the plan by calling the next() method on the next operator down. The ‘Root’ operator drives the control flow for the whole plan. Data is moved row by row throughout the entire plan; inefficient in terms of CPU instructions per row and prone to expensive CPU cache misses.
- #13: According to the Microsoft research paper on improvements made to column store indexes and batch mode in SQL 2014, the large object cache is new and it stores blob data contiguously in memory without any page breaks. The reason for this is that sequential page access for a CPU cache gives twice the throughput compared to single page access. Refer to slide 73 of “Super Scaling SQL Server Diagnosing and Fixing Hard Problems” by Thomas Kejser.
- #15: Not much difference, why ?. Column stores do not store data in sorted order, however the encoding and compression process can reorder data in order to help achieve better levels of compression.
- #17: Delta stores are new to SQL Server 2014 and they provide the means via which existing column stores can be inserted into. SQL Server 2014 also introduces column store archive compression. Writes to blobs, which row groups are stored as, are sequential in nature, for trickle inserts the presence of a delta store (b-tree) to act as a buffer mitigates against this. Updates take place by the deletion bit map for the column store being set and a new row being inserted into the column store via a delta store.