Lecture2-MapReduce - An introductory lecture to Map Reduce
- 1. Tutorial Meeting #2 Data Science and Machine Learning Map-Reduce and the New Software Stack Περικλής Ανδρίτσος
- 2. Introduction 2 • Once upon a time…. • There used to be just a single computer • It used to perform calculations all by itself • Later in the years to come…. • Data sources started proliferating • Size of the data sources that are used in calculations exploded • Advanced programming techniques started to appear • That’s when advanced science came in and gave us “distributed computation” • MapReduce is a representative of such technology
- 3. MapReduce (section 2) 3 • 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
- 4. Single Node Architecture (section 2.1) 4 Memory Disk CPU Machine Learning, Statistics “Classical” Data Mining { Main needs: a storage infrastructure and a programming paradigm (for analysis)
- 5. Motivation: Google Example 5 • 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, both in terms of storing it as well as the time needed to analyze it !! • Today, a standard architecture for such problems is emerging: • Cluster of commodity Linux nodes • Commodity network (ethernet) to connect them
- 6. Big picture of the new paradigm 6 • Split the data into chunks • Have multiple disks and CPUs • Distribute the data chunks across multiple disks • Process the data in parallel across CPUs • Example: • Time to read data: 4million secs (46+ days) • If we had 1000 CPUs we can do the task in • 4 million / 1000 = 4000 seconds (about 1 hour)
- 7. Cluster Architecture (Section 2.1.1) 7 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 It has been estimated that Google has 1M machines, http://bit.ly/Shh0RO
- 8. 8
- 9. Large-scale Computing 9 • 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 lose 1/day • If Google has ~1M machines, then • 1,000 machines fail every day!
- 10. Large-scale Computing (2) 10 • Problem #1: If nodes fail at such a rate, how do we make sure data is stored PERSISTENTLY • i.e., we are guaranteed to have data (or its copies) if machines that store it fail. • What if nodes fail during a long-running process what do we do ? • E.g., do we restart the process all over again ? • Solution: a new infrastructure that ”hides” all these failures
- 11. Large-scale Computing (3) 11 • Problem #2: Network Bottleneck • If bandwidth = 1 Gbps, moving 10TB takes ~1 day • In typical applications data moves around to be analyzed • In our case we should be moving enormous amounts of data into thousands of servers and that can be prohibitive • Solution: a new framework that doesn’t move data around ! • In my opinion: the big beauty of MapReduce !!
- 12. Map-Reduce (Section 2.2) 12 • It addresses previously mentioned problems in the following ways: • Stores data redundantly on multiple nodes (computers) for persistence and availability • Move computation close to the data to minimize data movement. This is a simple yet powerful idea. • Simple programming model that even non-savvy programmers can implement
- 13. Idea and Solution 13 • Issue: Copying data over a network takes time • Idea: • Bring computation close to the data • Store files multiple times for reliability • Map-reduce addresses these problems • Google’s computational/data manipulation model • Elegant way to work with big data • Storage Infrastructure – File system • Google: GFS. Hadoop: HDFS • Programming model • Map-Reduce
- 14. Storage Infrastructure 14 • Problem: • If nodes fail, how to store data persistently? • Answer: • Distributed File System (redundant storage): • Provides global file namespace • Implementations: Google GFS; Hadoop HDFS; • Stores each data piece multiple times across servers • Typical usage pattern • Huge files (100s of GB to TB) • Data is rarely updated in place (e.g. storing URLs) • Reads and appends are common
- 15. Distributed File System 15 • 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 D1 C5 C1 C3 C5 … C2 D0 D0 Bring computation directly to the data! C0 C5 Chunk server 1 Chunk server 2 Chunk server 3 Chunk server N C2 D0 Chunk servers also serve as compute servers
- 16. Distributed File System 16 • 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 • Master node • a.k.a. Name Node in Hadoop’s HDFS • Stores information 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
- 17. Coordination: Master 17 • 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
- 18. Dealing with Failures 18 • 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 • Only in-progress tasks are reset to idle • Reduce task is restarted • Master failure • MapReduce task is aborted and client is notified
- 19. Fault tolerance: Handled via re-execution • On worker failure: • Detect failure via periodic heartbeats • Re-execute completed and in-progress map tasks • Task completion committed through master • Robust: Google lost 1600 of 1800 machines, but kept running fine
- 20. Programming Model: MapReduce 20 Warm-up task: • We have a huge text document • Maybe 10-20TB of size • Count the number of times each distinct word appears in the file • Sample application: • Analyze web server logs to find popular URLs • Keyword statistic for search
- 21. Task: Word Count 21 Case 1: • File too large for memory, but all <word, count> pairs fit in memory <- simple program ! Case 2: • 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
- 22. MapReduce: Overview 22 words(doc.txt) | sort • Map • Scan input file one record at a time • Extract something you care about from each record (keys) • Group by key • Groups all the keys with the same value • Sort and shuffle • Reduce • Run a function • Aggregate, summarize, filter or transform • Write the answer
- 23. MapReduce: The Map Step 24 v k map v k v k … map Input key-value pairs Intermediate key-value pairs k v k v k v … k v
- 24. MapReduce: The Reduce Step 25 k v … Intermediate key-value pairs k v k v k v Group by key reduce reduce k v … k v … v Key-value groups k v v k v v Output key-value pairs k v k v
- 25. More Specifically 26 • 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’ • Note: * means a set
- 26. Map-Reduce: Environment 27 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
- 27. Map-Reduce: A diagram (centralized) 28 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 Map function Reduce functio n
- 28. Map-Reduce: In Parallel 29 Nod e Ensures all list with same key are sent to the same reduce node All phases are distributed with many
- 29. Map-Reduce 30 • 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
- 30. Data Flow 31 • 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 • Output is often input to another MapReduce task
- 31. MapReduce: Word Counting 32 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 - 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 Provided by the programmer (key, value) (key, value) O S n e l q y ue s n e ti q a u l l e y n r t e i a l d th r e e a d a s ta Key appears programmer 1 time
- 32. Word Count Using MapReduce 33 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; values: an iterator over counts result = 0 for each count v in values: result += v emit(key, result)
- 33. Refinement: Partition Function 36 • 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
- 34. Selection by Map-Reduce • Given a table R compute the operation 𝜎𝑐𝑜𝑛𝑑𝑖𝑡𝑖𝑜𝑛 𝑅 , where condition is a logical expression. Selection chooses the tuples from R that satisfy the condition. • Alternatively, in SQL, the operation is • SELECT attributes • FROM R • WHERE “condition is true” • 𝑹 37 A B a1: b1 a2 b1 a3 b2 a4 b3 1 𝜎𝐵=𝑏 (𝑅) A B a1: b1 a2 b1
- 35. Selection by Map-Reduce 38 • Mapper • Emit only tuples that satisfy the selection condition • The (key,value) pairs are of the form (t,t), where t is a tuples satisfying the condition. • In the previous example ((a1,b1),(a1,b1)) ((a2,b1),(a2,b1)) • Reducer • No reducer is required
- 36. GROUP BY and AGGREGATION by Map-Reduce 39 • Compute the query, Q: SELECT A,B,SUM(C) FROM R GROUP BY A,B • R(A,B,C) is stored in a file A B C A1 b1 1 A1 b1 2 A3 b2 3 A3 b2 4 R A B C A1 b1 3 A3 b2 7 � � Α, B, SUM(C) (R)
- 37. GROUP BY and AGGREGATION by Map-Reduce 40 • Mapper • The (Key, Value) pairs are of the form: • Key =<attributes in GROUP BY> • Value =<attributes in AGGREGATION. • In the previous example ((A1,B1),1) ((A1,B1),2) ((A3,B2),3) ((A3,B2),4) • Reducer • Applies the aggregation operation on the list of values and outputs results • In the previous example (A1,B1,3) (A3,B2,7)
- 38. Natural Join By Map-Reduce 41 • Compute the natural join R(A,B) ⋈ S(B,C) • R and S are each stored in files • Tuples are pairs (a,b) or (b,c) A B a1 b1 a2 b1 a3 b2 a4 b3 B C b2 c1 b2 c2 b3 c3 ⋈ A B C a3 b2 c1 a3 b2 c2 a4 b3 c3 = R S R⋈S
- 39. Map-Reduce Join 42 • Use a hash function h from B-values to 1...k • A Map process turns: • Each input tuple R(a,b) into key-value pair (b,(a,R)) • Each input tuple S(b,c) into (b,(c,S)) • Map processes send each key-value pair with key b to Reduce process h(b) • Hadoop does this automatically; just tell it what k is. • Each Reduce process matches all the pairs (b,(a,R)) with all (b,(c,S)) and outputs (a,b,c).
- 40. Matrix-Vector Multiplication with Map/Reduce Task: Compute product C = A·B 1 0 0 3 4 5 0 6 A 2 20 B 0 110 120 C 20 X =
- 41. X Computing Sparse Matrix-Vector Product 〈 2, 1, 10, B 〉 〈 3, 1, 20, B 〉 ● Represent matrix as list of nonzero entries 〈 row, col, value, matrixID 〉 ● Strategy ○ Phase 1: Compute all products ai,k · bk,j ○ Phase 2: Sum products for each entry i,j ○ Each phase involves a Map/Reduce 0 20 B 1 〈 1, 1, 1, A 〉 0 〈 1, 2, 2, A 〉 〈 2, 2, 3, A 〉 〈 2, 3, 4, A 〉 〈 3, 1, 5, A 〉 〈 3, 3, 6, A 〉 0 3 4 5 0 6 A 2
- 42. Group by - Map of Matrix-Vector Multiply Key = 2 Key = 3 〈 row, col, value, matrixID 〉→ (col, (matrixID, row, value)) 〈 1, 1, 1, A 〉→ (1, (A, 1, 1)) 〈 1, 2, 2, A 〉→ (2, (A, 1, 2)) 〈 2, 2, 3, A 〉→ (2, (A, 2, 3)) 〈 2, 3, 4, A 〉→ (3, (A, 2, 4)) 〈 3, 1, 5, A 〉→ (1, (A, 3, 5)) 〈 3, 3, 6, A 〉→ (3, (A, 3, 6)) Key = col 〈 2, 1, 10, B 〉→ (2, (B, 1, 10)) 〈 3, 1, 20, B 〉→ (3, (B, 1, 20)) Key = row Key = 1 (1, (A, 1, 1)) (1, (A, 3, 5)) (2, (A, 1, 2)) (2, (A, 2, 3)) (2, (B, 1, 10)) (3, (A, 2, 4)) (3, (A, 3, 6)) (3, (B, 1, 20)) Group By Intermediate Keys Mapping of Initial representation to intermediate keys 1 0 0 3 4 5 0 6 A 2 20 Group values ai,k and bk,j according to key k B 0
- 43. Group by - Reduce of Matrix-Vector Multiply Generate all products ai,k · bk,j X X 〈 1, 1, 2x10=20, C 〉 〈 2, 1, 3x10=30, C 〉 〈 2, 1, 4x20=80, C 〉 〈 3, 1, 6x20=120, C 〉 Key = 1 (1, (A, 1, 1)) (1, (A, 3, 5)) (2, (A, 1, 2)) Key = 2 (2, (A, 2, 3)) X (2, (B, 1, 10)) (3, (B, 1, 20)) Key = 3 (3, (A, 2, 4)) (3, (A, 3, 6)) 1 0 0 3 4 5 0 6 A 2 20 B 0 X
- 44. Aggregate - Map of Matrix-Vector Multiply Group products ai,k · bk,j with matching values of i and j Key = 1,1 Key = 2,1 Key = 3,1 〈 2, 1, 30, C 〉→ ((2, 1), (C, 30)) 〈 2, 1, 80, C 〉→ ((2, 1), (C, 80)) 〈 3, 1, 120, C 〉→ ((3, 1), (C, 120)) Key = (row,col) ((2, 1), (C, 30)) ((2, 1), (C, 80)) ((1, 1), (C, 20)) ((3, 1), (C, 120)) Mapping of intermediate representation to intermediate keys 〈 row, col, value, matrixID 〉→ ((row, col), (matrixID, value)) 〈 1, 1, 20, C 〉→ ((1, 1), (C, 20)) 1 0 0 3 4 5 0 6 A 2 20 B 0 X
- 45. Aggregate - Reduce of Matrix-Vector Multiply Sum products to get final entries 20 110 120 C 〈 1, 1, 20, C 〉 〈 2, 1, 110, C 〉 〈 3, 1, 120, C 〉 Key = 2,1 ((2, 1), (C, 30)) ((2, 1), (C, 80)) Key = 1,1 ((1, 1), (C, 20)) + + Key = 3,1 ((3, 1), (C, 120))+ 1 0 0 3 4 5 0 6 A 2 20 B 0 X
- 46. Cost Measures for Algorithms 49 • In MapReduce we quantify the cost of an algorithm using 1. Communication cost = total I/O of all processes 2. Computation cost analogous, but count only running time of processes Note that here the big-O notation is not the most useful (adding more machines is always an option)
- 47. Example: Hosting size 50 • 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 • Map • For each record, output (hostname(URL), size) • Reduce • Sum up the sizes of each host
- 48. Example: Language Model 51 • 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