Hw09 Rethinking The Data Warehouse With Hadoop And Hive
Download as PPT, PDF10 likes1,368 views
The document discusses Hive, a system for managing and querying large datasets stored in Hadoop. It describes how Hive provides a familiar SQL-like interface, simplifying Hadoop programming. The document also outlines how Facebook uses Hive and Hadoop for analytics, with over 4TB of new data added daily across a large cluster.
Hw09 Rethinking The Data Warehouse With Hadoop And Hive
- 1. Rethinking Data Warehousing & Analytics Ashish Thusoo, Facebook Data Infrastructure Team
- 2. Why Another Data Warehousing System? Data, data and more data 200GB per day in March 2008 2+TB(compressed) raw data per day in April 2009 4+TB(compressed) raw data per day today
- 4. Trends Leading to More Data Free or low cost of user services Realization that more insights are derived from simple algorithms on more data
- 5. Deficiencies of Existing Technologies Cost of Analysis and Storage on proprietary systems does not support trends towards more data Closed and Proprietary Systems Limited Scalability does not support trends towards more data
- 6. Hadoop Advantages Pros Superior in availability/scalability/manageability despite lower single node performance Open system Scalable costs Cons: Programmability and Metadata Map-reduce hard to program (users know sql/bash/python/perl) Need to publish data in well known schemas Solution: HIVE
- 7. What is HIVE? A system for managing and querying structured data built on top of Hadoop Components Map-Reduce for execution HDFS for storage Metadata in an RDBMS
- 8. Hive: Simplifying Hadoop – New Technology Familiar Interfaces hive> select key, count(1) from kv1 where key > 100 group by key; vs. $ cat > /tmp/reducer.sh uniq -c | awk '{print $2"\t"$1}‘ $ cat > /tmp/map.sh awk -F '\001' '{if($1 > 100) print $1}‘ $ bin/hadoop jar contrib/hadoop-0.19.2-dev-streaming.jar -input /user/hive/warehouse/kv1 -mapper map.sh -file /tmp/reducer.sh -file /tmp/map.sh -reducer reducer.sh -output /tmp/largekey -numReduceTasks 1 $ bin/hadoop dfs –cat /tmp/largekey/part*
- 9. Hive: Open and Extensible Query your own formats and types with your own Serializer/Deserializers Extend the SQL functionality through User Defined Functions Do any non-SQL transformations through TRANSFORM operator that sends data from Hive to any user program/script
- 10. Hive: Smart Execution Plans for Performance Hash based Aggregations Map-side Joins Predicate Pushdown Partition Pruning Many more to come in the future
- 11. Interoperability JDBC and ODBC interfaces available Integrations with some traditional SQL tools with some minor modifications More improvements in future to support interoperability with existing front end tools
- 12. Information Available as a sub project in Hadoop http://wiki.apache.org/hadoop/Hive (wiki) http://hadoop.apache.org/hive (home page) http://svn.apache.org/repos/asf/hadoop/hive (SVN repo) ##hive (IRC) Works with hadoop-0.17, 0.18, 0.19, 0.20 Release 0.4.0 is coming in the next few days Mailing Lists: hive-{user,dev,commits}@hadoop.apache.org
- 13. Data Warehousing @ Facebook using Hive & Hadoop
- 14. Data Flow Architecture at Facebook Web Servers Scribe MidTier Filers Production Hive-Hadoop Cluster Oracle RAC Federated MySQL Scribe-Hadoop Cluster Adhoc Hive-Hadoop Cluster Hive replication
- 15. Looks like this .. Disks Node Disks Node Disks Node Disks Node Disks Node Disks Node 1 Gigabit 4 Gigabit Node = DataNode + Map-Reduce
- 16. Hadoop & Hive Cluster @ Facebook Hadoop/Hive Warehouse – the new generation 4800 cores, Storage capacity of 5.5 PetaBytes 12 TB per node Two level network topology 1 Gbit/sec from node to rack switch 4 Gbit/sec to top level rack switch
- 17. Hive & Hadoop Usage @ Facebook Statistics per day: 4 TB of compressed new data added per day 135TB of compressed data scanned per day 7500+ Hive jobs on per day 80K compute hours per day Hive simplifies Hadoop: New engineers go though a Hive training session ~200 people/month run jobs on Hadoop/Hive Analysts (non-engineers) use Hadoop through Hive 95% of jobs are Hive Jobs
- 18. Hive & Hadoop Usage @ Facebook Types of Applications: Reporting Eg: Daily/Weekly aggregations of impression/click counts Measures of user engagement Microstrategy dashboards Ad hoc Analysis Eg: how many group admins broken down by state/country Machine Learning (Assembling training data) Ad Optimization Eg: User Engagement as a function of user attributes Many others
- 19. Facebook’s contributions… A lot of significant contributions: Hive Hdfs features Scheduler work Etc… Talks by Dhruba Borthakur and Zheng Shao in the development track for more information on these projects
Editor's Notes
- #3: Cost of training people is high – have to reduce cost by making system easy to use.
- #4: Why Hive? Petabytes of structured data User base familiar with SQL and Python/Perl/PHP Commercial Warehousing Software .. Does not scale, very expensive, inflexible Closed source, not programmable using Python/Perl/PHP Solution: SQL layer on top of scalable storage and map-reduce (Hadoop) Openness: Use any data format, embed any programming language
- #16: Nomenclature: Core switch and Top of Rack
- #18: 1GB connectivity within a rack, 100MB across racks? Are all disks 7200 SATA?