Elephant in the cloud

94
Elephant in the Cloud: a quest for the next generation Hadoop architecture Roman Shaposhnik Sr. Manager, Open Source Hadoop Platform @Pivotal (Twitter: @rhatr)

Transcript of Elephant in the cloud

Page 1: Elephant in the cloud

Elephant in the Cloud: a quest for the next generation

Hadoop architecture

Roman Shaposhnik Sr. Manager, Open Source Hadoop Platform @Pivotal

(Twitter: @rhatr)

Page 2: Elephant in the cloud

Who’s this guy?

•  Sr. Manager @Pivotal building a team of OS contributors

•  Apache Software Foundation guy (VP of Apache Incubator, VP of Apache Bigtop, committer on Hadoop, Giraph, Sqoop, etc)

•  Used to be root@Cloudera

•  Used to be PHB@Yahoo! (original Hadoop team)

•  Used to be a hacker at Sun microsystems (Sun Studio compilers and tools)

Page 3: Elephant in the cloud

Agenda

&

Page 4: Elephant in the cloud

Agenda

Page 5: Elephant in the cloud

Long, long time ago…

HDFS

ASF Projects FLOSS Projects Pivotal Products

MapReduce

Page 6: Elephant in the cloud

In a blink of an eye:

HDFS

Pig

Sqoop Flume

Coordination and workflow

management

Zookeeper

Command Center

ASF Projects FLOSS Projects Pivotal Products

GemFire XD

Oozie

MapReduce

Hive

Tez

Giraph

Hadoop UI

Hue

SolrCloud

Phoenix

HBase

Crunch Mahout

Spark

Shark

Streaming

MLib

GraphX

Impala

HAW

Q

SpringXD

MADlib

Ham

ster

PivotalR

YARN

Page 7: Elephant in the cloud

Genesis of Hadoop

• Google papers on GFS and MapReduce

• A subproject of Apache Nutch

• A bet by Yahoo!

Page 8: Elephant in the cloud

Data brings value

• What features to add to the product

• Data analysis must enable decisions

• V3: volume, velocity, variety

Page 9: Elephant in the cloud

Big Data brings big value

Page 10: Elephant in the cloud

Entering: Industrial Data

Page 11: Elephant in the cloud

Big Data Utility Gap

70% of data generated by

customers

80% of data being stored

3% being prepared for

analysis

0.5% being analyzed

<0.5% being operationalized

Average Enterprises

3 Exabytes per day

now

40 Trillion total Gigabytes in 2020

(Or 162 iPhones of storage for every

human)

?

Page 12: Elephant in the cloud
Page 13: Elephant in the cloud

Hadoop’s childhood

• HDFS: Hadoop Distributed Filesystem

• MapReduce: computational framework

Page 14: Elephant in the cloud
Page 15: Elephant in the cloud

HDFS: not a POSIXfs • Huge blocks: 64Mb (128Mb)

• Mostly immutable files (append, truncate)

• Streaming data access

• Block replication

Page 16: Elephant in the cloud

How do I use it?

$ hadoop fs –lsr / # hadoop-fuse-dfs dfs://hadoop-hdfs /mnt $ ls /mnt # mount –t nfs –o vers=3,proto=tcp,nolock host:/ /mnt $ ls /mnt

Page 17: Elephant in the cloud

Principle #1

HDFS is the datalake

Page 18: Elephant in the cloud

Pivotal’s Focus on Data Lakes

Existing EDW / Datamarts

Raw “untouched” Data

In-Mem

ory Parallel Ingest

Data Management���

(Search Engine)

Processed Data

In-Memory Services BI / A

nalytical Tools

Data Lake

ERP

HR

SFDC

New Data Sources/Formats

Machine

Traditional Data Sources

Finally! I now have full

transparency on the data

with amazing speed!

All data��� is now

accessible!

I can now afford ���“Big Data”

Business Users

ELT Processing with Hadoop

HDFS MapReduce/SQL/Pig/Hive

Analytical Data Marts/

Sandboxes

Security and Control

Page 19: Elephant in the cloud

HDFS enables the stack

HDFS

Pig

Sqoop Flume

Coordination and workflow

management

Zookeeper

Command Center

ASF Projects FLOSS Projects Pivotal Products

GemFire XD

Oozie

MapReduce

Hive

Tez

Giraph

Hadoop UI

Hue

SolrCloud

Phoenix

HBase

Crunch Mahout

Spark

Shark

Streaming

MLib

GraphX

Impala

HAW

Q

SpringXD

MADlib

Ham

ster

PivotalR

YARN

Page 20: Elephant in the cloud

Principle #2

Apps share their internal state

Page 21: Elephant in the cloud

MapReduce

• Batch oriented (long jobs; final results)

• Brings the computation to the data

• Very constrained programming model

• Embarrassingly parallel programming model

• Used to be the only game in town for compute

Page 22: Elephant in the cloud

MapReduce Overview

• Record = (Key, Value)

• Key : Comparable, Serializable

• Value: Serializable

• Logical Phases: Input, Map, Shuffle, Reduce, Output

Page 23: Elephant in the cloud

Map

• Input: (Key1, Value1)

• Output: List(Key2, Value2)

• Projections, Filtering, Transformation

Page 24: Elephant in the cloud

Shuffle

• Input: List(Key2, Value2)

• Output

• Sort(Partition(List(Key2, List(Value2))))

• Provided by Hadoop : Several Customizations Possible

Page 25: Elephant in the cloud

Reduce

• Input: List(Key2, List(Value2))

• Output: List(Key3, Value3)

• Aggregations

Page 26: Elephant in the cloud

Anatomy of MapReduce

d a c

a b c

a 3 b 1 c 2

a 1 b 1 c 1

a 1 c 1 a 1

a 1 1 1 b 1 c 1 1

HDFS mappers reducers HDFS

Page 27: Elephant in the cloud

MapReduce DataFlow

Page 28: Elephant in the cloud

How do I use it? public static class TokenizerMapper extends Mapper<Object, Text, Text, IntWritable> { public void map(Object key, Text value, Context context) { StringTokenizer itr = new StringTokenizer(value.toString()); while (itr.hasMoreTokens()) { word.set(itr.nextToken()); context.write(word, one); } } }

Page 29: Elephant in the cloud

How do I use it? public static class IntSumReducer extends Reducer<Text,IntWritable,Text,IntWritable> { public void reduce(Text key, Iterable<IntWritable> values, Context context) { int sum = 0; for (IntWritable val : values) { sum += val.get(); } result.set(sum); context.write(key, result); } }

Page 30: Elephant in the cloud

How do I run it?

$ hadoop jar hadoop-examples.jar wordcount \ input \ output

Page 31: Elephant in the cloud

Principle #3

MapReduce is assembly language of Hadoop

Page 32: Elephant in the cloud

Hadoop’s childhood

• Compact (pretty much a single jar)

• Challenged in scalability and SPOFs

• Extremely batch oriented

• Hard for non-Java programmers

Page 33: Elephant in the cloud

Then, something happened

Page 34: Elephant in the cloud

Hadoop 1.0

HDFS

ASF Projects FLOSS Projects Pivotal Products

MapReduce

Page 35: Elephant in the cloud

Hadoop 2.0

HDFS

ASF Projects FLOSS Projects Pivotal Products

MapReduce Tez

YARN

Ham

ster

YARN

Page 36: Elephant in the cloud

Hadoop 2.0

• HDFS 2.0

• Yet Another Resource Negotiator (YARN)

• MapReduce is just an “application” now

• Tez is another “application”

• Pivotal’s Hamster (OpenMPI) yet another one

Page 37: Elephant in the cloud

MapReduce 1.0

Job Tracker

Task Tracker���(HDFS)

Task Tracker���(HDFS)

task1 task1 task1 task1 task1

task1 task1 task1 task1 taskN

Page 38: Elephant in the cloud

YARN (AKA MR2.0)

Resource���Manager

Job Tracker

task1 task1 task1 task1 task1 Task Tracker

Page 39: Elephant in the cloud

YARN (AKA MR2.0)

Resource���Manager

Job Tracker

task1 task1 task1 task1 task1 Task Tracker

Page 40: Elephant in the cloud

YARN • Yet Another Resource Negotiator

• Resource Manager

• Node Managers

• Application Masters

• Specific to paradigm, e.g. MR Application master (aka JobTracker)

Page 41: Elephant in the cloud

YARN: beyond MR

Resource���Manager

MPI

MPI

Page 42: Elephant in the cloud

Hamster

•  Hadoop and MPI on the same cluster

•  OpenMPI Runtime on Hadoop YARN

•  Hadoop Provides: Resource Scheduling, ���Process monitoring, Distributed File System

•  Open MPI Provides: Process launching, ���Communication, I/O forwarding

Page 43: Elephant in the cloud

Hamster Components • Hamster Application Master

• Gang Scheduler, YARN Application Preemption

• Resource Isolation (lxc Containers)

• ORTE: Hamster Runtime

• Process launching, Wireup, Interconnect

Page 44: Elephant in the cloud

Hamster Architecture

Page 45: Elephant in the cloud

Hadoop 2.0

HDFS

ASF Projects FLOSS Projects Pivotal Products

MapReduce Tez

YARN

Ham

ster

YARN

Page 46: Elephant in the cloud

Hadoop ecosystem

HDFS

Pig

Sqoop Flume

Coordination and workflow

management

Zookeeper

Command Center

ASF Projects FLOSS Projects Pivotal Products

Oozie

MapReduce

Hive

Tez

Giraph

Hadoop UI

Hue

SolrCloud

Phoenix

HBase

Crunch Mahout

YARN

Ham

ster

YARN

Page 47: Elephant in the cloud

There’s way too much stuff

• Tracking dependencies

• Integration testing

• Optimizing the defaults

• Rationalizing the behaviour

Page 48: Elephant in the cloud

Wait! We’ve seen this!

GNU Software Linux kernel

Page 49: Elephant in the cloud

Apache Bigtop Hadoop ecosystem (Hbase, Pig, Hive)

Hadoop���(HDFS, YARN, MR)

Page 50: Elephant in the cloud

Principle #4

Apache Bigtop is how the Hadoop distros get

defined

Page 51: Elephant in the cloud

The ecosystem • Apache HBase

• Apache Crunch, Pig, Hive and Phoenix

• Apache Giraph

• Apache Oozie

• Apache Mahout

• Apache Sqoop and Flume

Page 52: Elephant in the cloud

Apache HBase • Small mutable records vs. HDFS files

• HFiles kept in HDFS

• Memcached for HDFS

• Built on HDFS and Zookeeper

• Google’s Bigtable

Page 53: Elephant in the cloud

Hbase datamodel

• Driven by the original Webtable usecase:

com.cnn.www <html>...

content:

CNN CNN.co

anchor:a.com anchor:b.com

Page 54: Elephant in the cloud

How do I use it? HTable table = new HTable(config, “table”);

Put p = new Put(Bytes.toBytes(“row”));

p.add(Bytes.toBytes(“family”),

Bytes.toBytes(“qualifier”),

Bytes.toBytes(“data”));

table.put(p);

Page 55: Elephant in the cloud

Dataflow model

HBase

HDFS

Producer Consumer

Page 56: Elephant in the cloud

When do I use it?

• Serving up large amounts of data

• Fast random access

• Scan operations

Page 57: Elephant in the cloud

Principle #5

HBase: when you need OLAP + OLTP

Page 58: Elephant in the cloud

What if its OLTP?

HDFS

Pig

Sqoop Flume

Coordination and workflow

management

Zookeeper

Command Center

ASF Projects FLOSS Projects Pivotal Products

Oozie

MapReduce

Hive

Tez

Giraph

Hadoop UI

Hue

SolrCloud

Phoenix

HBase

Crunch Mahout

YARN

Ham

ster

YARN

Page 59: Elephant in the cloud

GemFire XD

HDFS

Pig

Sqoop Flume

Coordination and workflow

management

Zookeeper

Command Center

ASF Projects FLOSS Projects Pivotal Products

Oozie

MapReduce

Hive

Tez

Giraph

Hadoop UI

Hue

SolrCloud

Phoenix

HBase

Crunch Mahout

YARN

GemFire XD

Ham

ster

YARN

Page 60: Elephant in the cloud

GemFire XD: a better HBase? • Close sourced but extremely mature

• SQL/Objects/JSON data model

• High concurrency, high update load

• Mostly selective point queries (no scans)

• Tiered storage architecture

Page 61: Elephant in the cloud

YCSB Benchmark; Throughput is 2-12X

0

100000

200000

300000

400000

500000

600000

700000

800000

AU BU CU D FU LOAD

Th

rou

ghp

ut

(op

s/se

c)

HBase

4

8

12

16

0

100000

200000

300000

400000

500000

600000

700000

800000

AU BU CU D FU LOAD

Th

rou

ghp

ut

(op

s/se

c)

GemFire XD

4

8

12

16

Page 62: Elephant in the cloud

YCSB Benchmark; Latency is 2X – 20X better

0

2000

4000

6000

8000

10000

12000

14000

Lat

en

cy (μ

sec)

HBase

4

8

12

16

0

2000

4000

6000

8000

10000

12000

14000

Lat

en

cy (μ

sec)

GemFire XD

4

8

12

16

Page 63: Elephant in the cloud

Principle #6

There are always 3 implementations

Page 64: Elephant in the cloud

Querying data

• MapReduce: “an assembly language”

• Apache Pig: a data manipulation DSL (now Turing complete!)

• Apache Hive: a batch-oriented SQL on top of Hadoop

Page 65: Elephant in the cloud

How do I use Pig?

grunt> A = load ‘./input.txt’;

grunt> B = foreach A generate ��� flatten(TOKENIZE((chararray)$0)) as��� words;

grunt> C = group B by word;

grunt> D = foreach C generate COUNT(B), ��� group;

Page 66: Elephant in the cloud

How do I use Hive? CREATE TABLE docs (line STRING);

LOAD DATA INPATH 'text' OVERWRITE INTO TABLE docs;

CREATE TABLE word_counts AS

SELECT word, count(1) AS count FROM

(SELECT explode(split(line, '\s')) AS word FROM docs)

GROUP BY word

ORDER BY word;

Page 67: Elephant in the cloud

Can we short Oracle now?

• No indexing

• Batch oriented scheduling

• Optimization for long running queries

• Metadata management is still in flux

Page 68: Elephant in the cloud

[Close to] real-time SQL

• Impala (inspired by Google’s F1)

• Hive/Tez (AKA Stinger)

• Facebook’s Presto (Hive’s lineage)

• Pivotal’s HAWQ

Page 69: Elephant in the cloud

HAWQ

• GreenPlum MPP database core

• True ANSI SQL support

• HDFS storage backend

• Parquet support is coming

Page 70: Elephant in the cloud

Principle #7

SQL on Hadoop

Page 71: Elephant in the cloud

Feeding the elephant

Page 72: Elephant in the cloud

Getting data in: Flume • Designed for collecting log data

• Flexible deployment topology

Page 73: Elephant in the cloud

Sqoop: RDBMs connection • Sqoop 1

• A MapReduce tool

• Must use Oozie for workflows

• Sqoop 2

• Well, 0.99.x really

• A standalone service

Page 74: Elephant in the cloud

Spring XD

• Unified, distributed, extensible system for data ingestions, real time analytics and data exports

• Apache Licensed, not ASF

• A runtime service, not a library

• AKA “Oozie + Flume + Sqoop + Morphlines”

Page 75: Elephant in the cloud

How do I use it?

# deployment: ./xd-singlenode

$ ./xd-shell

xd:> hadoop config fs –namenode hdfs://nn:8020

xd:> stream create –definition “time | hdfs” ��� –name ticktock

xd:> stream destroy –name ticktock

Page 76: Elephant in the cloud

Feeding the Elephant

HDFS

Pig

Sqoop Flume

Coordination and workflow

management

Zookeeper

Command Center

ASF Projects FLOSS Projects Pivotal Products

Oozie

MapReduce

Hive

Tez

Giraph

Hadoop UI

Hue

SolrCloud

Phoenix

HBase

Crunch Mahout

YARN

GemFire XD

SpringXD

Ham

ster

YARN

Page 77: Elephant in the cloud

Spark the disruptor

HDFS

Pig

Sqoop Flume

Coordination and workflow

management

Zookeeper

Command Center

ASF Projects FLOSS Projects Pivotal Products

GemFireXD

Oozie

MapReduce

Hive

Tez

Giraph

Hadoop UI

Hue

SolrCloud

Phoenix

HBase

Crunch Mahout

Spark

Shark

Streaming

MLib

GraphX

SpringXD

YARN

Ham

ster

YARN

Page 78: Elephant in the cloud

What’s wrong with MR?

Source: UC Berkeley Spark project (just the image)

Page 79: Elephant in the cloud

Spark innovations • Resilient Distribtued Datasets (RDDs)

• Distributed on a cluster

• Manipulated via parallel operators (map, etc.)

• Automatically rebuilt on failure

• A parallel ecosystem

• A solution to iterative and multi-stage apps

Page 80: Elephant in the cloud

RDDs

warnings = textFile(…).filter(_.contains(“warning”)) .map(_.split(‘ ‘)(1))

HadoopRDD���path = hdfs://

FilteredRDD���contains…

MappedRDD split…

Page 81: Elephant in the cloud

Parallel operators

• map, reduce

• sample, filter

• groupBy, reduceByKey

• join, leftOuterJoin, rightOuterJoin

• union, cross

Page 82: Elephant in the cloud

An alternative backend

• Shark: a Hive on Spark

• Spork: a Pig on Spark

• Mlib: machine learning on Spark

• GraphX: Graph processing on Spark

• Also featuring its own streaming engine

Page 83: Elephant in the cloud

How do I use it?

val file = spark.textFile("hdfs://...")

val counts = file.flatMap(line => line.split(" "))

.map(word => (word, 1))

.reduceByKey(_ + _)

counts.saveAsTextFile("hdfs://...")

Page 84: Elephant in the cloud

Principle #8

Spark is the technology of 2014

Page 85: Elephant in the cloud

Where’s the cloud?

Page 86: Elephant in the cloud

What’s new?

• True elasticity

• Resource partitioning

• Security

• Data marketplace

• Data leaks/breaches

Page 87: Elephant in the cloud

Hadoop Maturity

ETL Offload Accommodate massive ���

data growth with existing EDW investments

Data Lakes Unify Unstructured and Structured Data Access

Big Data Apps

Build analytic-led applications impacting ���

top line revenue

Data-Driven Enterprise

App Dev and Operational Management on HDFS

Data Architecture

Page 88: Elephant in the cloud

Pivotal HD on Pivotal CF

� Enterprise PaaS Management System

� Flexible multi-language ‘buildpack’ architecture

� Deployed applications enjoy built-in services

� On-Premise Hadoop as a Service

� Single cluster deployment of Pivotal HD

� Developers instantly bind to shared Hadoop Clusters

� Speeds up time-to-value

Page 89: Elephant in the cloud

Pivotal Data Fabric Evolution

Analytic���Data Marts

SQL Services

Operational ���Intelligence

In-Memory Database

Run-Time���Applications

Data Staging���Platform

Data Mgmt. Services

Pivotal Data Platform

Stream ���Ingestion

Streaming Services

Software-Defined Datacenter

New Data-fabrics

In-Memory Grid

...ETC

Page 90: Elephant in the cloud

Principle #9

Hadoop in the Cloud is one of many

distributed frameworks

Page 91: Elephant in the cloud

2014 is the year of Hadoop

HDFS

Pig

Sqoop Flume

Coordination and workflow

management

Zookeeper

Command Center

ASF Projects FLOSS Projects Pivotal Products

GemFire XD

Oozie

MapReduce

Hive

Tez

Giraph

Hadoop UI

Hue

SolrCloud

Phoenix

HBase

Crunch Mahout

Spark

Shark

Streaming

MLib

GraphX

Impala

HAW

Q

SpringXD

MADlib

Ham

ster

PivotalR

YARN

Page 92: Elephant in the cloud

A NEW PLATFORM FOR A NEW ERA

Additional Line 18 Point Verdana

Page 93: Elephant in the cloud

Credits

• Apache Software Foundation

• Milind Bhandarkar

• Konstantin Boudnik

• Robert Geiger

• Susheel Kaushik

• Mak Gokhale

Page 94: Elephant in the cloud

Questions ?