ch02-mapreduce.pptx

 Challenges:
 How to distribute computation?
 Distributed/parallel programming is hard
 Map-reduce addresses all of the above
 Google’s computational/data manipulation model
 Elegant way to work with big data
2

Memory
Disk
CPU
Machine Learning, Statistics
“Classical” Data Mining
3

 20+ billion web pages x 20KB = 400+ TB
 1 computer reads 30-35 MB/sec from disk
 ~4 months to read the web
 ~1,000 hard drives to store the web
 Takes even more to do something useful
with the data!
 Today, a standard architecture for such
problems is emerging:
 Cluster of commodity Linux nodes
 Commodity network (ethernet) to connect them
4

Mem
Disk
CPU
Mem
Disk
CPU
…
Switch
Each rack contains 16-64 nodes
Mem
Disk
CPU
Mem
Disk
CPU
…
Switch
Switch
1 Gbps between
any pair of nodes
in a rack
2-10 Gbps backbone between racks
In 2011 it was guestimated that Google had 1M machines, http://bit.ly/Shh0RO
5

J. Leskovec, A. Rajaraman, J. Ullman: Mining of Massive Datasets, http://www.mmds.org 6

 Large-scale computing for data mining
problems on commodity hardware
 Challenges:
 How do you distribute computation?
 How can we make it easy to write distributed
programs?
 Machines fail:
 One server may stay up 3 years (1,000 days)
 If you have 1,000 servers, expect to loose 1/day
 People estimated Google had ~1M machines in 2011
 1,000 machines fail every day!
7

 Large-scale computing for data mining
problems on commodity hardware
 Challenges:
 How do you distribute computation?
 How can we make it easy to write distributed
programs?
 Machines fail:
 One server may stay up 3 years (1,000 days)
 If you have 1,000 servers, expect to loose 1/day
 People estimated Google had ~1M machines in 2011
 1,000 machines fail every day!
8

 Problem:
 If nodes fail, how to store data persistently?
 Answer:
 Distributed File System:
 Provides global file namespace
 Google GFS; Hadoop HDFS;
 Typical usage pattern
 Huge files (100s of GB to TB)
 Data is rarely updated in place
 Reads and appends are common
9

 Reliable distributed file system
 Data kept in “chunks” spread across machines
 Each chunk replicated on different machines
 Seamless recovery from disk or machine failure
C0 C1
C2
C5
Chunk server 1
D1
C5
Chunk server 3
C1
C3
C5
Chunk server 2
…
C2
D0
D0
Bring computation directly to the data!
C0 C5
Chunk server N
C2
D0
10
Chunk servers also serve as compute servers

 Chunk servers
 File is split into contiguous chunks
 Typically each chunk is 16-64MB
 Each chunk replicated (usually 2x or 3x)
 Try to keep replicas in different racks (why do that?)
 Master node
 a.k.a. Name Node in Hadoop’s HDFS
 Stores metadata about where files are stored
 Might be replicated
 Client library for file access
 Talks to master to find chunk servers
 Connects directly to chunk servers to access data
11

Warm-up task:
 We have a huge text document
 Count the number of times each
distinct word appears in the file
 Sample application:
 Analyze web server logs to find popular URLs
 Term statistics for search
12

Case 1:
 File too large for memory, but all <word, count> pairs
fit in memory (solution HashTable)
Case 2:
 File too large for memory and all <word, count> pairs
do not fit in memory (need Big Data Soution)
 Count occurrences of words:
 words(doc.txt) | sort | uniq -c
 where words takes a file and outputs the words in it, one per
a line
 Case 2 captures the essence of MapReduce
 Great thing is that it is naturally parallelizable
13

 Sequentially read a lot of data
 Map: (words(doc.txt))
 Extract something you care about e.g. word=key
 Group by key: (sort)
 Sort and Shuffle
 Reduce: (uniq –c)
 Aggregate, summarize, filter or transform
 Write the result
Outline stays the same, Map and Reduce
change to fit the problem
14

v
k
k v
k v
map
v
k
v
k
…
k v
map
Input
key-value pairs
Intermediate
key-value pairs
…
k v
15

k v
…
k v
k v
k v
Intermediate
key-value pairs
Group
by key
reduce
reduce
k v
k v
k v
…
k v
…
k v
k v v
v v
Key-value groups
Output
key-value pairs
16

 Input: a set of key-value pairs
 Programmer specifies two methods:
 Map(k, v)  <k’, v’>*
 Takes a key-value pair and outputs a set of key-value pairs
 E.g., key is the filename, value is a single line in the file
 There is one Map call for every (k,v) pair
 Reduce(k’, <v’>*)  <k’, v’’>*
 All values v’ with same key k’ are reduced together
and processed in v’ order
 There is one Reduce function call per unique key k’
17

The crew of the space
shuttle Endeavor recently
returned to Earth as
ambassadors, harbingers of
a new era of space
exploration. Scientists at
NASA are saying that the
recent assembly of the
Dextre bot is the first step in
a long-term space-based
man/mache partnership.
'"The work we're doing now
-- the robotics we're doing -
- is what we're going to
need ……………………..
Big document
(The, 1)
(crew, 1)
(of, 1)
(the, 1)
(space, 1)
(shuttle, 1)
(Endeavor, 1)
(recently, 1)
….
(crew, 1)
(crew, 1)
(space, 1)
(the, 1)
(the, 1)
(the, 1)
(shuttle, 1)
(recently, 1)
…
(crew, 2)
(space, 1)
(the, 3)
(shuttle, 1)
(recently, 1)
…
MAP:
Read input and
produces a set of
key-value pairs
Group by key:
Collect all pairs
with same key
Reduce:
Collect all values
belonging to the
key and output
(key, value)
Provided by the
programmer
Provided by the
programmer
(key, value)
(key, value)
Sequentially
read
the
data
Only
sequential
reads
18

map(key, value):
// key: document name; value: text of the document
for each word w in value:
emit(w, 1)
reduce(key, values):
// key: a word; value: an iterator over counts
result = 0
for each count v in values:
result += v
emit(key, result)
19

21
Big document
MAP:
Read input and
produces a set of
key-value pairs
Group by key:
Collect all pairs with
same key
(Hash merge, Shuffle,
Sort, Partition)
Reduce:
Collect all values
belonging to the
key and output

22
Use a Partitioning (HASH) function in order to find which
reducer e.g. key k4 is send to Reducer1
The programmer supplies the function R to do something
with the values.
All phases are distributed with many tasks doing the work

Map-Reduce environment takes care of:
 Partitioning the input data
 Scheduling the program’s execution across a
set of machines
 Performing the group by key step
 Handling machine failures
 Managing required inter-machine
communication
23

 Programmer specifies:
 Map and Reduce and input files
 Workflow:
 Read inputs as a set of key-value-
pairs
 Map transforms input kv-pairs into a
new set of k'v'-pairs
 Sorts & Shuffles the k'v'-pairs to
output nodes
 All k’v’-pairs with a given k’ are sent
to the same reduce
 Reduce processes all k'v'-pairs
grouped by key into new k''v''-pairs
 Write the resulting pairs to files
 All phases are distributed with
many tasks doing the work
Input 0
Map 0
Input 1
Map 1
Input 2
Map 2
Reduce 0 Reduce 1
Out 0 Out 1
Shuffle
24

 Input and final output are stored on a
distributed file system (FS):
 Scheduler tries to schedule map tasks “close” to
physical storage location of input data
 Intermediate results are stored on local FS
of Map and Reduce workers e.g. output of a
map step
 Output of a MapReduce Task is often input
to another MapReduce task
25

 Master node takes care of coordination:
 Task status: (idle, in-progress, completed)
 Idle tasks get scheduled as workers become
available
 When a map task completes, it sends the master
the location and sizes of its R intermediate files,
one for each reducer
 Master pushes this info to reducers
 Master pings workers periodically to detect
failures
26

 Map worker failure
 Map tasks completed or in-progress at
worker are reset to idle
 Reduce workers are notified when task is
rescheduled on another worker
 Reduce worker failure
 Reduce task is restarted
 Master failure
 MapReduce task is aborted and client is notified
27

 M map tasks, R reduce tasks
 Rule of a thumb:
 Make M much larger than the number of nodes
in the cluster
 One DFS chunk per map is common
 Improves dynamic load balancing and speeds up
recovery from worker failures
 Usually R is smaller than M
 Because output is spread across R files (more
convenient to store a file/output in a small
number of files/reducers)
28

 Often a Map task will produce many pairs of
the form (k,v1), (k,v2), … for the same key k
 E.g., popular words (e.g. the) in the word count
example
 Can save network time by
pre-aggregating values in
the mapper:
 E.g. Send (‘the’,1013)
 INPUT:combine(k, list(v1))
OUTPUT k,v2
 Combiner is usually same
as the reduce function
30

 Back to our word counting example:
 Combiner combines the values of all keys of a
single mapper (single machine):
 Much less data needs to be copied and shuffled!
31

 Want to control how keys get partitioned
 Inputs to map tasks are created by contiguous
splits of input file
 Reduce needs to ensure that records with the
same intermediate key end up at the same worker
 System uses a default partition function:
 hash(key) mod R
 Sometimes useful to override the hash
function:
 E.g., hash(hostname(URL)) mod R ensures URLs
from a host end up in the same output file
32

 Google
 Uses Google File System (GFS)
 Not available outside Google
 Hadoop
 An open-source implementation in Java
 Uses HDFS for stable storage
 Download: http://lucene.apache.org/hadoop/
 Aster Data
 Cluster-optimized SQL Database that also
implements MapReduce
33

 Ability to rent computing by the hour
 Additional services e.g., persistent storage
 Amazon’s “Elastic Compute Cloud” (EC2)
 Aster Data and Hadoop can both be run on
EC2
 For CS341 (offered next quarter) Amazon will
provide free access for the class
34

 Statistical machine translation:
 Need to count number of times every 5-word
sequence occurs in a large corpus of documents
 Very easy with MapReduce:
 Map:
 Extract (5-word sequence, count) from document
 Reduce:
 Combine the counts
36

 Suppose we have a large web corpus
 Look at the metadata file
 Lines of the form: (URL, size, date, …)
 For each host, find the total number of bytes
 That is, the sum of the page sizes for all URLs from
that particular host
 Other examples:
 Link analysis and graph processing
 Machine Learning algorithms
37

 Jeffrey Dean and Sanjay Ghemawat:
MapReduce: Simplified Data Processing on
Large Clusters
 http://labs.google.com/papers/mapreduce.html
 Sanjay Ghemawat, Howard Gobioff, and Shun-
Tak Leung: The Google File System
 http://labs.google.com/papers/gfs.html
39

 Hadoop Wiki
 Introduction
 http://wiki.apache.org/lucene-hadoop/
 Getting Started
 http://wiki.apache.org/lucene-
hadoop/GettingStartedWithHadoop
 Map/Reduce Overview
 http://wiki.apache.org/lucene-hadoop/HadoopMapReduce
 http://wiki.apache.org/lucene-
hadoop/HadoopMapRedClasses
 Eclipse Environment
 http://wiki.apache.org/lucene-hadoop/EclipseEnvironment
 Javadoc
 http://lucene.apache.org/hadoop/docs/api/
40

 Releases from Apache download mirrors
 http://www.apache.org/dyn/closer.cgi/lucene/had
oop/
 Nightly builds of source
 http://people.apache.org/dist/lucene/hadoop/nig
htly/
 Source code from subversion
 http://lucene.apache.org/hadoop/version_control
.html
41

 Programming model inspired by functional language
primitives
 Partitioning/shuffling similar to many large-scale sorting
systems
 NOW-Sort ['97]
 Re-execution for fault tolerance
 BAD-FS ['04] and TACC ['97]
 Locality optimization has parallels with Active
Disks/Diamond work
 Active Disks ['01], Diamond ['04]
 Backup tasks similar to Eager Scheduling in Charlotte
system
 Charlotte ['96]
 Dynamic load balancing solves similar problem as River's
distributed queues
 River ['99]
42

ch02-mapreduce.pptx

More Related Content

ch02-mapreduce.pptx