A common mistake size communication trade off specific tradeoffs
This presentation is the property of its rightful owner.
Sponsored Links
1 / 63

A Common Mistake Size/Communication Trade-Off Specific Tradeoffs PowerPoint PPT Presentation


  • 76 Views
  • Uploaded on
  • Presentation posted in: General

Designing Efficient Map-Reduce Algorithms. A Common Mistake Size/Communication Trade-Off Specific Tradeoffs. Jeffrey D. Ullman Stanford University. Research Is Joint Work of. Foto Afrati (NTUA) Anish Das Sarma (Google) Semih Salihoglu (Stanford) U. Motivating Example.

Download Presentation

A Common Mistake Size/Communication Trade-Off Specific Tradeoffs

An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -

Presentation Transcript


A common mistake size communication trade off specific tradeoffs

Designing Efficient Map-Reduce Algorithms

A Common MistakeSize/Communication Trade-OffSpecific Tradeoffs

Jeffrey D. Ullman

Stanford University


Research is joint work of

Research Is Joint Work of

  • FotoAfrati (NTUA)

  • Anish Das Sarma (Google)

  • SemihSalihoglu (Stanford)

  • U.


The drug interaction problem a failed attempt lowering the communication

Motivating Example

The Drug Interaction ProblemA Failed AttemptLowering the Communication


The drug interaction problem

The Drug-Interaction Problem

  • Data consists of records for 3000 drugs.

    • List of patients taking, dates, diagnoses.

    • About 1M of data per drug.

  • Problem is to find drug interactions.

    • Example: two drugs that when taken together increase the risk of heart attack.

  • Must examine each pair of drugs and compare their data.


Initial map reduce algorithm

Initial Map-Reduce Algorithm

  • The first attempt used the following plan:

    • Key = set of two drugs {i, j}.

    • Value = the record for one of these drugs.

  • Given drug i and its record Ri, the mapper generates all key-value pairs ({i, j}, Ri), where j is any other drug besides i.

  • Each reducer receives its key and a list of the two records for that pair: ({i, j}, [Ri,Rj]).


Example three drugs

Example: Three Drugs

Mapper

for drug 1

Reducer

for {1,2}

Drug 3 data

Drug 2 data

Drug 2 data

Drug 3 data

Drug 1 data

Drug 1 data

{2, 3}

{1, 3}

{2, 3}

{1, 2}

{1, 3}

{1, 2}

Mapper

for drug 2

Reducer

for {1,3}

Mapper

for drug 3

Reducer

for {2,3}


Example three drugs1

Example: Three Drugs

Mapper

for drug 1

Reducer

for {1,2}

Drug 2 data

Drug 3 data

Drug 2 data

Drug 3 data

Drug 1 data

Drug 1 data

{2, 3}

{1, 3}

{2, 3}

{1, 2}

{1, 3}

{1, 2}

Mapper

for drug 2

Reducer

for {1,3}

Mapper

for drug 3

Reducer

for {2,3}


Example three drugs2

Example: Three Drugs

Reducer

for {1,2}

{1, 2}

Drug 1 data

Drug 2 data

Reducer

for {1,3}

Drug 1 data

Drug 3 data

{1, 3}

Reducer

for {2,3}

Drug 2 data

Drug 3 data

{2, 3}


What went wrong

What Went Wrong?

  • 3000 drugs

  • times 2999 key-value pairs per drug

  • times 1,000,000 bytes per key-value pair

  • = 9 terabytes communicated over a 1Gb Ethernet

  • = 90,000 seconds of network use.


The improved algorithm

The Improved Algorithm

  • They grouped the drugs into 30 groups of 100 drugs each.

    • Say G1 = drugs 1-100, G2 = drugs 101-200,…, G30 = drugs 2901-3000.

    • Let g(i) = the number of the group into which drug i goes.


The map function

The Map Function

  • A key is a set of two group numbers.

  • The mapper for drug i produces 29 key-value pairs.

    • Each key is the set containing g(i) and one of the other group numbers.

    • The value is a pair consisting of the drug number i and the megabyte-long record for drug i.


The reduce function

The Reduce Function

  • The reducer for pair of groups {m, n} gets that key and a list of 200 drug records – the drugs belonging to groups m and n.

  • Its job is to compare each record from group m with each record from group n.

    • Special case: also compare records in group n with each other, if m = n+1 or if n = 30 and m = 1.

  • Notice each pair of records is compared at exactly one reducer, so the total computation is not increased.


The new communication cost

The New Communication Cost

  • The big difference is in the communication requirement.

  • Now, each of 3000 drugs’ 1MB records is replicated 29 times.

    • Communication cost = 87GB, vs. 9TB.


Reducer size replication rate mapping schemas lower bounds

Theory of Map-Reduce Algorithms

Reducer SizeReplication RateMapping SchemasLower Bounds


A model for map reduce algorithms

A Model for Map-Reduce Algorithms

  • A set of inputs.

    • Example: the drug records.

  • A set of outputs.

    • Example: One output for each pair of drugs.

  • A many-many relationship between each output and the inputs needed to compute it.

    • Example: The output for the pair of drugs {i, j} is related to inputs i and j.


Example drug inputs outputs

Example: Drug Inputs/Outputs

Output 1-2

Drug 1

Output 1-3

Drug 2

Output 1-4

Drug 3

Output 2-3

Drug 4

Output 2-4

Output 3-4


Example matrix multiplication

Example: Matrix Multiplication

j

j

i

i

=


Reducer size

Reducer Size

  • Reducer size, denoted q, is the maximum number of inputs that a given reducer can have.

    • I.e., the length of the value list.

  • Limit might be based on how many inputs can be handled in main memory.

  • Or: make q low to force lots of parallelism.


Replication rate

Replication Rate

  • The average number of key-value pairs created by each mapper is the replication rate.

    • Denoted r.

  • Represents the communication cost per input.


Example drug interaction

Example: Drug Interaction

  • Suppose we use g groups and d drugs.

  • A reducer needs two groups, so q = 2d/g.

  • Each of the d inputs is sent to g-1 reducers, or approximately r = g.

  • Replace g by r in q = 2d/g to get r = 2d/q.

Tradeoff!

The bigger the reducers,

the less communication.


Upper and lower bounds on r

Upper and Lower Bounds on r

  • What we did gives an upper bound on r as a function of q.

  • A solid investigation of map-reduce algorithms for a problem includes lower bounds.

    • Proofs that you cannot have lower r for a given q.


Proofs need mapping schemas

Proofs Need Mapping Schemas

  • A mapping schema for a problem and a reducer size q is an assignment of inputs to sets of reducers, with two conditions:

    • No reducer is assigned more than q inputs.

    • For every output, there is some reducer that receives all of the inputs associated with that output.

      • Say the reducer covers the output.


Mapping schemas 2

Mapping Schemas – (2)

  • Every map-reduce algorithm has a mapping schema.

  • The requirement that there be a mapping schema is what distinguishes map-reduce algorithms from general parallel algorithms.


Example drug interactions

Example: Drug Interactions

  • d drugs, reducer size q.

  • Each drug has to meet each of the d-1 other drugs at some reducer.

  • If a drug is sent to a reducer, then at most q-1 other drugs are there.

  • Thus, each drug is sent to at least (d-1)/(q-1) reducers, and r>(d-1)/(q-1).

  • Half the r from the algorithm we described.

  • Better algorithm gives r = d/q + 1, so lower bound is actually tight.


The better algorithm

The Better Algorithm

  • The problem with the algorithm dividing inputs into g groups is that members of a group appear together at many reducers.

    • Thus, each reducer can only productively compare about half the elements it gets.

  • Better: use smaller groups, with each reducer getting many little groups.

    • Eliminates almost all the redundancy.


Optimal algorithm for all pairs

Optimal Algorithm for All-Pairs

  • Assume d inputs.

  • Let p be a prime, where p2 divides d.

  • Divide inputs into p2 groups of d/p2 inputs each.

  • Name the groups (i, j), where 0 <i, j < p.

  • Use p(p+1) reducers, organized into p+1 teams of p reducers each.

  • For 0 < k < p, group (i, j) is sent to the reducer i+kj (mod p) in group k.

  • In the last team (p), group (i, j) is sent to reducer j.


Example teams for p 5

Example: Teams for p = 5

j = 0

1

2

3

4

i = 0

1

2

3

4

Team 0


Example teams for p 51

Example: Teams for p = 5

j = 0

1

2

3

4

i = 0

1

2

3

4

Team 1


Example teams for p 52

Example: Teams for p = 5

j = 0

1

2

3

4

i = 0

1

2

3

4

Team 2


Example teams for p 53

Example: Teams for p = 5

j = 0

1

2

3

4

i = 0

1

2

3

4

Team 3


Example teams for p 54

Example: Teams for p = 5

j = 0

1

2

3

4

i = 0

1

2

3

4

Team 4


Example teams for p 55

Example: Teams for p = 5

j = 0

1

2

3

4

i = 0

1

2

3

4

Team 5


Why it works

Why It Works

  • Let two inputs be in groups (i, j) and (i’, j’).

  • If the same group, these inputs obviously share a reducer.

  • If j = j’, then they share a reducer in team p.

  • If j  j’, then they share a reducer in team k provided i + kj = i’ + kj’ (all arithmetic modulo p).

  • Equivalently, (i-i’) = k(j-j’).

  • But since j  j’, (j-j’) has an inverse modulo p.

  • Thus, team k = (i-i’)(j-j’)-1 has a reducer for which i + kj = i’ + kj’.


Why it is optimal

Why It Is Optimal

  • The replication rate r is p+1, since every input is sent to one reducer in each team.

  • The reducer size q= p(d/p2) = d/p, since each reducer gets p groups of size d/p2.

  • Thus, r = d/q + 1.

  • (d/q + 1) - (d-1)/(q-1) < 1 provided q < d.

    • But if q> d, we can do everything in one reducer, and r = 1.

  • The upper bound r< d/q+ 1 and the lower bound r>(d-1)/(q-1) differ by less than 1, and are integers, so they are equal.


The exact lower bound matching algorithms

The Hamming-Distance = 1 Problem

The Exact Lower BoundMatching Algorithms


Definition of hd1 problem

Definition of HD1 Problem

  • Given a set of bit strings of length b, find all those that differ in exactly one bit.

  • Example: For b=2, the inputs are 00, 01, 10, 11, and the outputs are (00,01), (00,10), (01,11), (10,11).

  • Theorem: r> b/log2q.

    • (Part of) the proof later.


Algorithm with q 2

Algorithm With q=2

  • We can use one reducer for every output.

  • Each input is sent to b reducers (so r = b).

  • Each reducer outputs its pair if both its inputs are present, otherwise, nothing.

  • Subtle point: if neither input for a reducer is present, then the reducer doesn’t really exist.


Algorithm with q 2 b

Algorithm with q = 2b

  • Alternatively, we can send all inputs to one reducer.

  • No replication (i.e., r = 1).

  • The lone reducer looks at all pairs of inputs that it receives.


Splitting algorithm

Splitting Algorithm

  • Assume b is even.

  • Two reducers for each string of length b/2.

    • Call them the left and right reducers for that string.

  • String w = xy, where |x| = |y| = b/2, goes to the left reducer for x and the right reducer for y.

  • If w and z differ in exactly one bit, then they will both be sent to the same left reducer (if they disagree in the right half) or to the same right reducer (if they disagree in the left half).

  • Thus, r = 2; q = 2b/2.


Proof that r b log 2 q

Proof That r> b/log2q

  • Lemma: A reducer of size q cannot cover more than (q/2)log2q outputs.

    • Induction on b; proof omitted.

  • (b/2)2b outputs must be covered.

  • There are at least p = (b/2)2b/((q/2)log2q) = (b/q)2b/log2q reducers.

  • Sum of inputs over all reducers >pq = b2b/log2q.

  • Replication rate r = pq/2b = b/log2q.

    • Omits possibility that smaller reducers help.


A common mistake size communication trade off specific tradeoffs

Algorithms Matching Lower Bound

Generalized Splitting

One reducer

for each output

b

Splitting

All inputs

to one

reducer

r = replication

rate

2

1

2b/2

2b

21

q = reducer

size


One job method two job method comparison

Matrix Multiplication

One-Job MethodTwo-Job MethodComparison


Matrix multiplication

Matrix Multiplication

  • Assume n  n matrices AB = C.

  • Aij is the element in row i and column j of matrix A.

    • Similarly for B and C.

  • Cik = jAij Bjk.

  • Output Cikdepends on the ith row of A, that is, Aijfor all j, and the kth column of B, that is, Bjkfor all j.


Computing one output value

Computing One Output Value

Column k

Row i

=

A

B

C


Reducers cover rectangles

Reducers Cover Rectangles

  • Important fact: If a reducer covers outputs Cik and Cfg, then it also covers Cig and Cfk.

  • Why? This reducer has all of rows i and f of A as inputs and also has all of columns k and g of B as inputs.

  • Thus, it has all the inputs it needs to cover Cigand Cfk.

  • Generalizing: Each reducer covers all the outputs in the “rectangle” defined by a set of rows and a set of columns of matrix C.


The responsibility of one reducer

The Responsibility of One Reducer


Upper bound on output size

Upper Bound on Output Size

  • If a reducer gets q inputs, it gets q/n rows or columns.

  • Maximize the number of outputs covered by making the input “square.”

    • I.e., #rows = #columns.

  • q/2n rows and q/2n columns yield q2/4n2 outputs covered.


Lower bound on replication rate

Lower Bound on Replication Rate

  • Total outputs = n2.

  • One reducer can cover at most q2/4n2outputs.

  • Therefore, 4n4/q2 reducers.

  • 4n4/q total inputs to all the reducers, divided by 2n2 total inputs = 2n2/q replication rate.

  • Example: If q = 2n2, one reducer suffices and the replication rate is r = 1.

  • Example: If q = 2n (minimum possible), then r = n.


Matching algorithm

Matching Algorithm

  • Divide rows of the first matrix into g groups of n/g rows each.

  • Also divide the columns of the second matrix into g groups of n/g columns each.

  • g2 reducers, each with q = 2n2/g inputs consisting of a group of rows and a group of columns.

  • r = g = 2n2/q.


Picture of one reducer

Picture of One Reducer

n/g

n/g

=


Two job map reduce algorithm

Two-Job Map-Reduce Algorithm

  • A better way: use two map-reduce jobs.

  • Job 1: Divide both input matrices into rectangles.

    • Reducer takes two rectangles and produces partial sums of certain outputs.

  • Job 2: Sum the partial sums.


Picture of first job

Picture of First Job

K

K

J

J

I

I

A

B

C

For i in I and k in K, contribution

is j in J Aij × Bjk


First job details

First Job – Details

  • Divide the rows of the first matrix A into g groups of n/g rows each.

  • Divide the columns of A into 2g groups of n/2g.

  • Divide the rows of the second matrix B into 2g groups of n/2g rows each.

  • Divide the columns of B into g groups of n/g.

  • Important point: the groups of columns for A and rows for B must have indices that match.


Reducers for first job

Reducers for First Job

  • Reducers correspond to an n/g by n/2g rectangle in A (with row indices I, column indices J) and an n/2g by n/g rectangle in B (with row indices J and column indices K).

    • Call this reducer (I,J,K).

    • Important point: there is one set of indices J that plays two roles.

      • Needed so only rectangles that need to be multiplied are given a reducer.


The reducer i j k

The Reducer (I,J,K)

K

K

J

n/2g

J

I

n/g

n/g

I

n/g

n/g

n/2g

A

B

C

2g reducers contribute to

this area, one for each J.


Job 1 details

Job 1: Details

  • Convention: i, j, k are individual rows and/or column numbers, which are members of groups I, J, and K, respectively.

  • Mappers Job 1:

    • Aij -> key = (I,J,K) for any group K; value = (A,i,j,Aij).

    • Bjk-> key = (I,J,K) for any group I; value = (B,j,k,Bjk).

  • Reducers Job 1: For key (I,J,K) produce xiJk = j in J Aij Bjk for all i in I and k in K.


Job 2 details

Job 2: Details

  • Mappers Job 2: xiJk -> key = (i,k), value = xiJk.

  • Reducers Job 2: For key (i,k), produce output Cik = J xiJk.


Comparison computation cost

Comparison: Computation Cost

  • The two methods (one or two map-reduce jobs) essentially do the same computation.

    • Every Aij is multiplied once with every Bjk.

    • All terms in the sum for Cik are added together somewhere, only once.

  • 2 jobs requires some extra overhead of task management.


Comparison communication cost

Comparison: Communication Cost

  • One-job method: r = 2n2/q; there are 2n2 inputs, so total communication = 4n4/q.

  • Two-job method with parameter g:

    Job 2: Communication = (2g)(n2/g2)(g2) = 2n2g.

Number of reducers

contributing to

each output

Number of output squares

Area of each square


Communication cost continued

Communication Cost – Continued

  • Job 1 communication:

    • 2n2 input elements.

    • Each generates g key-value pairs.

    • So another 2n2g.

    • Total communication = 4n2g.

  • Reducer size q = (2)(n2/2g2) = n2/g2.

    • So g = n/q.

    • Total communication = 4n3/q.

      • Compares favorably with 4n4/q for the one-job approach.


Summary

Summary

  • Represent problems by mapping schemas

  • Get upper bounds on number of covered outputs as a function of reducer size.

  • Turn these into lower bounds on replication rate as a function of reducer size.

  • For HD = 1 and all-pairs problems: exact match between upper and lower bounds.

  • 1-job matrix multiplication analyzed exactly.

  • But 2-job MM yields better total communication.


Research questions

Research Questions

  • Get matching upper and lower bounds for the Hamming-distance problem for distances greater than 1.

    • Ugly fact: For HD=1, you cannot have a large reducer with all pairs at distance 1; for HD=2, it is possible.

      • Consider all inputs of weight 1 and length b.


Research questions 2

Research Questions – (2)

  • Give an algorithm that takes an input-output mapping and a reducer size q, and gives a mapping schema with the smallest replication rate.

  • Is the problem even tractable?

  • A recent extension by Afrati, Dolev, Korach, Sharma, and U. lets inputs have weights, and the reducer size limits the sum of the weights of the inputs received.

    • What can be extended to this model?


  • Login