SQL on Hadoop - 12th Swiss Big Data User Group Meeting, 3rd of July, 2014, ETH Zurich
1 like1,172 views
This document discusses the evolution of SQL on Hadoop. There are two types of SQL engines on Hadoop - Type I engines that compile SQL to MapReduce, and Type II engines that use their own distributed execution engines. Early engines like Hive were Type I and had limitations due to MapReduce. Newer engines like Impala, Presto, Spark SQL, and IBM BigSQL are Type II and can perform SQL queries directly on Hadoop files without MapReduce. IBM BigSQL provides ANSI SQL with performance optimized for big data by leveraging decades of database query optimization experience.
![© 2013 IBM Corporation21
BigSQL V1.0 – Demo (small)
[bivm.ibm.com][biadmin] 1> select count(*) from trace1;
+----------+
| |
+----------+
| 11416740 |
+----------+
1 row in results(first row: 39.78s; total: 39.78s)](https://image.slidesharecdn.com/sqlonhadoopmasterbigdatausergroup-140703121229-phpapp02/85/SQL-on-Hadoop-12th-Swiss-Big-Data-User-Group-Meeting-3rd-of-July-2014-ETH-Zurich-21-320.jpg)
![© 2013 IBM Corporation23
BigSQL V1.0 – Demo (small)
[bivm.ibm.com][biadmin] 1> select count(*) from trace1 t3 inner
join trace2 t4 on t3.hour=t4.hour;
+--------+
| |
+--------+
| 477340 |
+--------+
1 row in results(first row: 32.24s; total: 32.25s)](https://image.slidesharecdn.com/sqlonhadoopmasterbigdatausergroup-140703121229-phpapp02/85/SQL-on-Hadoop-12th-Swiss-Big-Data-User-Group-Meeting-3rd-of-July-2014-ETH-Zurich-23-320.jpg)
![© 2013 IBM Corporation25
BigSQL V3.0 – Demo (small)
[bivm.ibm.com][biadmin] 1> select count(*) from trace3;
+----------+
| 1 |
+----------+
| 12014733 |
+----------+
1 row in results(first row: 2.94s; total: 2.95s)](https://image.slidesharecdn.com/sqlonhadoopmasterbigdatausergroup-140703121229-phpapp02/85/SQL-on-Hadoop-12th-Swiss-Big-Data-User-Group-Meeting-3rd-of-July-2014-ETH-Zurich-25-320.jpg)
![© 2013 IBM Corporation26
BigSQL V3.0 – Demo (small)
[bivm.ibm.com][biadmin] 1> select count(*) from trace3 t3 inner
join trace4 t4 on t3.hour=t4.hour;
+--------+
| 1 |
+--------+
| 504360 |
+--------+
1 row in results(first row: 0.79s; total: 0.80s)](https://image.slidesharecdn.com/sqlonhadoopmasterbigdatausergroup-140703121229-phpapp02/85/SQL-on-Hadoop-12th-Swiss-Big-Data-User-Group-Meeting-3rd-of-July-2014-ETH-Zurich-26-320.jpg)
![© 2013 IBM Corporation27
BigSQL V3.0 – Demo (small)
[bivm.ibm.com][biadmin] 1> select count(hour), hour from trace3
group by hour order by hour;
29 rows in results(first row: 1.88s; total: 1.89s)](https://image.slidesharecdn.com/sqlonhadoopmasterbigdatausergroup-140703121229-phpapp02/85/SQL-on-Hadoop-12th-Swiss-Big-Data-User-Group-Meeting-3rd-of-July-2014-ETH-Zurich-27-320.jpg)
SQL on Hadoop - 12th Swiss Big Data User Group Meeting, 3rd of July, 2014, ETH Zurich
- 1. © 2013 IBM Corporation1 SQL on Hadoop - 12th Swiss Big Data User Group Meeting, 3rd of July, 2014, ETH Zurich Romeo Kienzler IBM Center of Excellence for Data Science, Cognitive Systems and BigData (A joint-venture between IBM Research Zurich and IBM Innovation Center DACH) Source: http://www.kdnuggets.com/2012/04/data-science-history.jpg
- 2. © 2013 IBM Corporation2 DataScience at present ● Tools (http://blog.revolutionanalytics.com/2014/01/in-data-scientist-survey-r-is-the-most-used-tool-other-than-databases.html) ● SQL (42%) ● R (33%) ● Python (26%) ● Excel (25%) ● Java, Ruby, C++ (17%) ● SPSS, SAS (9%) ● Limitations (Single Node usage) ● Main Memory ● CPU <> Main Memory Bandwidth ● CPU ● Storage <> Main Memory Bandwidth (either Single node or SAN)
- 3. © 2013 IBM Corporation3 Data Science on Hadoop SQL (42%) R (33%) Python (26%) Excel (25%) Java, Ruby, C++ (17%) SPSS, SAS (9%) Data Science Hadoop
- 4. © 2013 IBM Corporation4 SQL on Hadoop ● IBM BigSQL (ANSI 2011 compliant, part of IBM BigInsights) ● HIVE, Presto ● Cloudera Impala ● Lingual ● Shark ● ... SQL Hadoop
- 5. © 2013 IBM Corporation5 Two types of SQL Engines ● Type I ● Compiler and Optimizer SQL->MapReduce ● Type II ● Brings own distributed execution engine on Data Nodes ● Brings own Task Scheduler ● The Hadoop SQL Ecosystem is evolving very fast
- 6. © 2013 IBM Corporation6 Hive ● Runs on top of MapReduce ● → Type I Source: http://cdn.venublog.com/wp-content/uploads/2013/07/hive-1.jpg
- 7. © 2013 IBM Corporation7 Lingual ● ANSI SQL Layer on top of Cascading ● Cascading ● Java API do express DAG ● Runs on top of MapReduce ● → Type I
- 8. © 2013 IBM Corporation8 Limits of MapReduce ● Disk writes between Map and Reduce ● Slow for computations which depend on previously computed values ● JOINs are very slow and difficult to implement ● Only sequential data access ● Only tuple-wise data access ● Map-Side joins have sort and size constraints ● Reduce-Side joins require secondary sorting of values ● … ● ...
- 9. © 2013 IBM Corporation9 Impala (Type II) http://blog.cloudera.com/blog/wp-content/uploads/2012/10/impala.png
- 10. © 2013 IBM Corporation10 Presto (Type II) https://www.facebook.com/notes/facebook-engineering/presto-interacting-with-petabytes-of-data-at-facebook/10151786197628920
- 11. © 2013 IBM Corporation11 Spark / Shark (Type II) Source: http://bighadoop.files.wordpress.com/2014/04/spark-architecture.png
- 12. © 2013 IBM Corporation12 BigSQL V3.0 (Type II) Like in Spark, MapReduce has been Kicked out :) (No JobTracker, No Task Tracker, But HDFS/GPFS remains)
- 13. © 2013 IBM Corporation13 BigSQL V3.0 – Architecture Putting the story together…. Big SQL shares a common SQL dialect with DB2 Big SQL shares the same client drivers with DB2
- 14. © 2013 IBM Corporation14 BigSQL V3.0 – Performance Query rewrites Exhaustive query rewrite capabilities Leverages additional metadata such as constraints and nullability Optimization Statistics and heuristic driven query optimization Query optimizer based upon decades of IBM RDBMS experience Tools and metrics Highly detailed explain plans and query diagnostic tools Extensive number of available performance metrics SELECT ITEM_DESC, SUM(QUANTITY_SOLD), AVG(PRICE), AVG(COST) FROM PERIOD, DAILY_SALES, PRODUCT, STORE WHERE PERIOD.PERKEY=DAILY_SALES.PERKEY AND PRODUCT.PRODKEY=DAILY_SALES.PRODKE Y AND STORE.STOREKEY=DAILY_SALES.STOREKEY AND CALENDAR_DATE BETWEEN AND '01/01/2012' AND '04/28/2012' AND STORE_NUMBER='03' AND CATEGORY=72 GROUP BY ITEM_DESC Access plan generationQuery transformation Dozens of query transformations Hundreds or thousands of access plan options Store Product Product Store NLJOIN Daily SalesNLJOIN Period NLJOIN Product NLJOIN Daily Sales NLJOIN Period NLJOIN Store HSJOIN Daily Sales HSJOIN Period HSJOIN Product StoreZZJOIN Daily Sales HSJOIN Period
- 15. © 2013 IBM Corporation15 BigSQL V3.0 – Performance You are substantially faster if you don't use MapReduce IBM BigInsights v3.0, with Big SQL 3.0, is the only Hadoop distribution to successfully run ALL 99 TPC-DS queries and ALL 22 TPC-H queries without modification. Source: http://www.ibmbigdatahub.com/blog/big-deal-about- infosphere-biginsights-v30-big-sql
- 16. © 2013 IBM Corporation16 BigSQL V3.0 – Query Federation Head Node Big SQL Compute Node Task Tracker Data Node Big SQL Compute Node Task Tracker Data Node Big SQL Compute Node Task Tracker Data Node Big SQL Compute Node Task Tracker Data Node Big SQL
- 17. © 2013 IBM Corporation17 BigSQL V1.0 – Demo (small) ● 32 GB Data, ~650.000.000 rows (small, Innovation Center Zurich) ● 3 TB Data, ~ 60.937.500.000 rows (middle, Innovation Center Zurich) ● 0.7 PB Data, ~ 1.421875×10¹³ rows (large, Innovation Center Hursley) ● 32 GB Data, ~650.000.000 rows (small, Innovation Center Zurich) ● 3 TB Data, ~ 60.937.500.000 rows (middle, Innovation Center Zurich) ● 0.7 PB Data, ~ 1.421875×10¹³ rows (large, Innovation Center Hursley)
- 18. © 2013 IBM Corporation18 BigSQL V1.0 – Demo (small) CREATE EXTERNAL TABLE trace ( hour integer, employeeid integer, departmentid integer, clientid integer, date string, timestamp string) ROW FORMAT DELIMITED FIELDS TERMINATED BY ',' LINES TERMINATED BY 'n' STORED AS TEXTFILE LOCATION '/user/biadmin/32Gtest';
- 19. © 2013 IBM Corporation19 BigSQL V1.0 – Demo (small)
- 20. © 2013 IBM Corporation20 BigSQL V1.0 – Demo (small)
- 21. © 2013 IBM Corporation21 BigSQL V1.0 – Demo (small) [bivm.ibm.com][biadmin] 1> select count(*) from trace1; +----------+ | | +----------+ | 11416740 | +----------+ 1 row in results(first row: 39.78s; total: 39.78s)
- 22. © 2013 IBM Corporation22 BigSQL V1.0 – Demo (small) select count(hour), hour from trace group by hour order by hour 30 rows in results(first row: 37.98s; total: 37.99s)
- 23. © 2013 IBM Corporation23 BigSQL V1.0 – Demo (small) [bivm.ibm.com][biadmin] 1> select count(*) from trace1 t3 inner join trace2 t4 on t3.hour=t4.hour; +--------+ | | +--------+ | 477340 | +--------+ 1 row in results(first row: 32.24s; total: 32.25s)
- 24. © 2013 IBM Corporation24 BigSQL V3.0 – Demo (small) CREATE HADOOP TABLE trace3 ( hour int, employeeid int, departmentid int,clientid int, date varchar(30), timestamp varchar(30) ) row format delimited fields terminated by '|' stored as textfile;
- 25. © 2013 IBM Corporation25 BigSQL V3.0 – Demo (small) [bivm.ibm.com][biadmin] 1> select count(*) from trace3; +----------+ | 1 | +----------+ | 12014733 | +----------+ 1 row in results(first row: 2.94s; total: 2.95s)
- 26. © 2013 IBM Corporation26 BigSQL V3.0 – Demo (small) [bivm.ibm.com][biadmin] 1> select count(*) from trace3 t3 inner join trace4 t4 on t3.hour=t4.hour; +--------+ | 1 | +--------+ | 504360 | +--------+ 1 row in results(first row: 0.79s; total: 0.80s)
- 27. © 2013 IBM Corporation27 BigSQL V3.0 – Demo (small) [bivm.ibm.com][biadmin] 1> select count(hour), hour from trace3 group by hour order by hour; 29 rows in results(first row: 1.88s; total: 1.89s)
- 28. © 2013 IBM Corporation28 Questions? http://www.ibm.com/software/data/bigdata/ BigInsights free VM and Installer for non-commercial use: ibm.co/quickstart Twitter: @RomeoKienzler, @IBMEcosystem_DE, @IBM_ISV_Alps