Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

46
Interactive Analytics With Spark And Cassandra Evan Chan Ooyala, Inc. April 7Th, 2014

description

This session covers our experience with using the Spark and Shark frameworks for running real-time queries on top of Cassandra data.We will start by surveying the current Cassandra analytics landscape, including Hadoop and HIVE, and touch on the use of custom input formats to extract data from Cassandra. We will then dive into Spark and Shark, two memory-based cluster computing frameworks, and how they enable often dramatic improvements in query speed and productivity, over the standard solutions today.

Transcript of Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Page 1: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Interactive Analytics With Spark And Cassandra

!Evan Chan

Ooyala, Inc. April 7Th, 2014

Page 2: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

• Staff Engineer, Compute and Data Services, Ooyala

• Building multiple web-scale real-time systems on top of C*, Kafka, Storm, etc.

• Scala/Akka guy

• Very excited by open source, big data projects

• @evanfchan

Who is this guy?

!2

Page 3: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

• Cassandra at Ooyala

• What problem are we trying to solve?

• Spark and Shark

• Integrating Cassandra and Spark

• Our Spark/Cassandra Architecture

Agenda

!3

Page 4: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

CASSANDRA AT OOYALA

!4

Page 5: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

OOYALA Powering personalized video

experiences across all screens.

Page 6: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

CONFIDENTIAL—DO NOT DISTRIBUTE !6CONFIDENTIAL—DO NOT DISTRIBUTE

Founded in 2007

Commercially launch in 2009

230+ employees in Silicon Valley, LA, NYC, London, Paris, Tokyo, Sydney & Guadalajara

Global footprint, 200M unique users,110+ countries, and more than 6,000 websites

Over 1 billion videos played per month and 2 billion analytic events per day

25% of U.S. online viewers watch video powered by Ooyala

COMPANY OVERVIEW

Page 7: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

CONFIDENTIAL—DO NOT DISTRIBUTE !7

TRUSTED VIDEO PARTNER

STRATEGIC PARTNERS

CUSTOMERS

CONFIDENTIAL—DO NOT DISTRIBUTE

Page 8: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

TITLE TEXT GOES HERE

• 12 clusters ranging in size from 3 to 107 nodes

• Total of 28TB of data managed over ~220 nodes

• Powers all of our analytics infrastructure

• Traditional analytics aggregations

• Recommendations and trends

• DSE/C* 1.0.x, 1.1.x, 1.2.6

We are a large Cassandra user

!8

Page 9: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

TITLE TEXT GOES HERE

• Started investing in Spark beginning of 2013

• 2 teams of developers doing stuff with Spark

• Actively contributing to Spark developer community

• Deploying Spark to a large (>100 node) production cluster

• Spark community very active, huge amount of interest

Becoming a big Spark user...

!9

Page 10: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

WHAT PROBLEM ARE WE TRYING TO SOLVE?

!10

Page 11: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

From mountains of raw data...

Page 12: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

• Quickly

• Painlessly

• At scale?

To nuggets of truth...

Page 13: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Today: Precomputed Aggregates

• Video metrics computed along several high cardinality dimensions

• Very fast lookups, but inflexible, and hard to change

• Most computed aggregates are never read

• What if we need more dynamic queries?

• Top content for mobile users in France

• Engagement curves for users who watched recommendations

• Data mining, trends, machine learning

Page 14: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

The Static - Dynamic Continuum

• Super fast lookups

• Inflexible, wasteful

• Best for 80% most common queries

• Always compute results from raw data

• Flexible but slow

100% Precomputation 100% Dynamic

Page 15: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Where We Want To Be

Partly dynamic

• Pre-aggregate most common queries

• Flexible, fast dynamic queries

• Easily generate many materialized

views

Page 16: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

WHY SPARK?

!16

Page 17: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Introduction To Spark• In-memory distributed computing framework

• Created by UC Berkeley AMP Lab in 2010

• Targeted problems that MR is bad at:

– Iterative algorithms (machine learning)

– Interactive data mining

• More general purpose than Hadoop MR

• Top level Apache project

• Active contributions from Intel, Yahoo, lots of

Page 18: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Spark Vs Hadoop

HDFS

Map

Reducee

Map

Reduce

Data  Source

map()

join()

Source  2

cache()

transform

Page 19: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Throughput: Memory Is King

6-node C*/DSE 1.1.9 cluster, Spark 0.7.0

Spark cached RDD 10-50x faster than raw Cassandra

Page 20: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Developers Love It

• “I wrote my first aggregation job in 30 minutes”

• High level “distributed collections” API

• No Hadoop cruft

• Full power of Scala, Java, Python

• Interactive REPL shell

Page 21: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Spark Vs Hadoop Word Count

file = spark.textFile("hdfs://...") file.flatMap(line => line.split(" "))    .map(word => (word, 1))    .reduceByKey(_ + _)

1 package org.myorg;! 2 ! 3 import java.io.IOException;! 4 import java.util.*;! 5 ! 6 import org.apache.hadoop.fs.Path;! 7 import org.apache.hadoop.conf.*;! 8 import org.apache.hadoop.io.*;! 9 import org.apache.hadoop.mapreduce.*;! 10 import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;! 11 import org.apache.hadoop.mapreduce.lib.input.TextInputFormat;! 12 import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;! 13 import org.apache.hadoop.mapreduce.lib.output.TextOutputFormat;! 14 ! 15 public class WordCount {! 16 ! 17 public static class Map extends Mapper<LongWritable, Text, Text, IntWritable> {! 18 private final static IntWritable one = new IntWritable(1);! 19 private Text word = new Text();! 20 ! 21 public void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {! 22 String line = value.toString();! 23 StringTokenizer tokenizer = new StringTokenizer(line);! 24 while (tokenizer.hasMoreTokens()) {! 25 word.set(tokenizer.nextToken());! 26 context.write(word, one);! 27 }! 28 }! 29 } ! 30 ! 31 public static class Reduce extends Reducer<Text, IntWritable, Text, IntWritable> {! 32 ! 33 public void reduce(Text key, Iterable<IntWritable> values, Context context) ! 34 throws IOException, InterruptedException {! 35 int sum = 0;! 36 for (IntWritable val : values) {! 37 sum += val.get();! 38 }! 39 context.write(key, new IntWritable(sum));! 40 }! 41 }! 42 ! 43 public static void main(String[] args) throws Exception {! 44 Configuration conf = new Configuration();! 45 ! 46 Job job = new Job(conf, "wordcount");! 47 ! 48 job.setOutputKeyClass(Text.class);! 49 job.setOutputValueClass(IntWritable.class);! 50 ! 51 job.setMapperClass(Map.class);! 52 job.setReducerClass(Reduce.class);! 53 ! 54 job.setInputFormatClass(TextInputFormat.class);! 55 job.setOutputFormatClass(TextOutputFormat.class);! 56 ! 57 FileInputFormat.addInputPath(job, new Path(args[0]));! 58 FileOutputFormat.setOutputPath(job, new Path(args[1]));! 59 ! 60 job.waitForCompletion(true);! 61 }! 62 ! 63 }!

Page 22: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

One Platform To Rule Them All

HIVE on Spark

Spark Streaming - discretized stream

processing

• SQL, Graph, ML, Streaming all in one framework

• Much higher code sharing/reuse • Easy integration between components • Fewer platforms == lower TCO • Integration with Mesos, YARN helps share resources

Page 23: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Shark - Hive On Spark

• 100% HiveQL compatible

• 10-100x faster than HIVE, answers in seconds

• Reuse UDFs, SerDe’s, StorageHandlers

• Can use DSE / CassandraFS for Metastore

Page 24: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

INTEGRATING CASSANDRA & SPARK

!24

Page 25: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Our Spark/Shark/Cassandra Stack

Node1

Cassandra

InputFormat

SerDe

Spark Worker

Shark

Node2

Cassandra

InputFormat

SerDe

Spark Worker

Shark

Node3

Cassandra

InputFormat

SerDe

Spark Worker

Shark

Spark Master Job Server

Page 26: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

OPTIONS FOR READING FROM C*

• Hadoop InputFormat

– ColumnFamilyInputFormat - reads all rows from 1 CF – CqlPagingInputFormat, etc. - CQL3, 2-dary indexes – Roll your own (join multiple CFs, etc)

• Spark native RDD – sc.parallelize(rowkeys).flatMap(readColumns(_))

– JdbcRdd + Cassandra JDBC driver

Page 27: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Columnfamilyinputformat

video type

Record1 10 1

Record2 11 5

id Video TypeRecord1 10 1

Record2 11 5

• Must read from all rows • One CF only, not very flexible

Page 28: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Node 2Node 1

Spark RDD• RDD = Resilient Distributed Dataset • Multiple partitions living on different nodes

• Each partition has records

Partition 1 Partition 2 Partition 3 Partition 4

Page 29: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Inputformat Vs Rdd

InputFormat RDDSupports Hadoop, HIVE, Spark, Shark

Spark / Shark only

Have to implement multiple classes - InputFormat, RecordReader, Writeable, etc. Clunky API.

One class - simple API.

Two APIs, and often need to implement both (HIVE needs older...)

Just one API.

• You can easily use InputFormats in Spark using newAPIHadoopRDD().

• Writing a custom RDD could have saved us lots of time.

Page 30: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Node 2Node 1

Skipping The Inputformat

Row 1 data Row 2 data Row 3 data Row 4 data

Rowkey1 Rowkey2 Rowkey3 Rowkey4

sc.parallelize(rowkeys).flatMap(readColumns(_))

Page 31: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

OUR CASSANDRA / SPARK ARCHITECTURE

!31

Page 32: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

From Raw Logs To Fast Queries

ProcessC*

columnar storage

Raw Logs

Raw Logs

Raw LogsSpark

Spark

Spark

OLAP Table 1

OLAP Table 2

OLAP Table 3

Spark

Shark

Predefined queries

Ad-hoc HiveQL

Page 33: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Why Cassandra Alone Isn’t Enough• Over 30 million multi-dimensional fact table rows per day

• Materializing every possible answer isn’t close to possible

• Multi dimensional filtering and grouping alone leads to many billions of possible answers

• Querying fact tables in Cassandra is too slow

• Reading millions or billion random rows

• CQL doesn’t support grouping

Page 34: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Our Approach

• Use Cassandra to store the raw fact tables

• Optimize the schema for OLAP workloads

• Fast full table reads • Easily read fewer columns

• Use Spark for fast random row access and fast distributed computation

Page 35: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

uuid-part-0

uuid-part-1

2013-04-05T00:00Z#id1

Section 0 Section 1 Section 2

country rows 0-9999 rows 10000-19999 rows ....

city rows 0-9999 rows 10000-19999 rows ....

Index CF

Columns CF

An OLAP Schema for Cassandra

Metadata

2013-04-05T00:00-part-0

{columns:[“country”, “city”, “plays” ]}

Metadata CF

•Optimized for: selective column loading, maximum throughput and compression

Page 36: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Olap Workflow

DatasetAggregation Job

Query JobSpark

Executors

Cassandra

REST Job Server

Query Job

Aggregate Query

Result

Query

Result

Page 37: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Querying Data In Spark

Node 2Node 1

Partition 1 !

record1 record2 record3

Partition 2 !

record4 record5 record6

Partition 3 !

record7 record8 record9

Partition 4 !

record10 record11 record12

Convert to Spark SQL / Shark Table

Shark / Spark SQL

rdd.group / .map / .sort / .join etc

rdd.reduce / .collect / .top

Page 38: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Fault Tolerance

• Cached dataset lives in Java Heap only - what if process dies?

• Spark lineage - automatic recomputation from source, but this is expensive!

• Can also replicate cached dataset to survive single node failures

• Persist materialized views back to C*, then load into cache -- now recovery path is much faster

Page 39: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

DEMO

!39

Page 40: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Creating a Shark Table

Page 41: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Creating a Cached Table

Page 42: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Querying a Cached Table

Page 43: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

THANK YOUAnd YES, We’re HIRING!!

ooyala.com/careers

@evanfchan

Page 44: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

Industry Trends

• Fast execution frameworks

• Impala

• In-memory databases

• VoltDB, Druid

• Streaming and real-time

• Higher-level, productive data

Page 45: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

PERFORMANCE #’S

Spark: C* -> OLAP aggregatescold cache, 1.4 million events

130 seconds

C* -> OLAP aggregateswarmed cache

20-30 seconds

OLAP aggregate query via Spark(56k records)

60 ms

6-node C*/DSE 1.1.9 cluster, Spark 0.7.0

Page 46: Cassandra Day SV 2014: Spark, Shark, and Apache Cassandra

EXAMPLE: OLAP PROCESSING

t02013-04-05T00:00Z#i

{video: 10,

2013-04-05T00:00Z#i

{video: 20,

C* events

OLAP Aggregates

OLAP Aggregates

OLAP Aggregates

Cached Materialized Views

Spark

Spark

Spark

Union

Query 1: Plays by Provider

Query 2: Top content for mobile