Intro to cassandra + hadoop
Download as PPTX, PDF4 likes3,247 views
This document introduces Cassandra and Hadoop and how they can be used together for analytics over Cassandra data. It discusses how Cassandra is good for writes and random reads at scale but not ad-hoc queries, while Hadoop tools like MapReduce, Pig, and Hive can query Cassandra data and are extensible. It provides examples of using MapReduce and Pig with Cassandra and discusses how Raptr.com uses Cassandra and Hadoop together to improve query performance from hours to 10-15 minutes.
Intro to cassandra + hadoop
- 1. Cassandra + HadoopAn Introduction to Hadoop Analytics over Cassandra Data
- 2. IntroductionsWhat is Cassandra?A highly scalable distributed data storeBorn at Facebook, grew up in the communityWhat is Hadoop?A set of Apache projectsDeal with Big Data in a distributed wayOpen source versions of MapReduce, GFS, BigTable, as well as additions, such as Pig and Hive
- 3. What makes them compatible?Cassandra is great at a lot of thingsFast, extremely scalable writes, fast random readsFlexible semi-structured data modelNot as good with ad-hoc answersEnter HadoopMapReduce, Pig, and Hive are extensibleOutput from Hadoop into Cassandra
- 4. MapReduceInput from Cassandra as of 0.6.xBaked in output to Cassandra as of 0.7.0Streaming support is coming in 0.7Example: WordCount
- 5. PigWhat is Pig?A platform for data analytics developed at Yahoo!Includes PigLatin, Grunt shell, and interpreter that compiles down to MapReduceSimplifies data analysisCassandra integrationStu Hood added Pig integration in Cassandra 0.6Example: WordCount with Pig
- 6. HiveWhat is Hive?A platform for data analytics developed at FacebookDraws from the familiar SQL -> Hive QLCompiles down to MapReduceCassandra integrationAvailability of a Cassandra storage handler is coming soon – HIVE-1434
- 7. Example Use CaseRaptr.comGaming statistics and achievements across platformsHome-grown -> Cassandra + Hadoop (Pig)Idea to execution much fasterQuery runtime from hours to 10-15 minutes
- 8. QuestionsContactEmail: [email protected]: @jeromatronIRC: jeromatron on irc.freenode.net - #cassandra, #hadoopFurther informationhttp://wiki.apache.org/cassandra/HadoopSupportCassandra: The Definitive Guide