Big Data Programming Map Reduce

Please answer this question. Please provide the description of your design, your source code, a screen shot showing the command you use to start your job, a screenshot showing results.

Here is the question (photos are attached and the matrix/code is also attached):

Recall in BDP-3-mapReduce-3 slides, we discussed the situation where vector v is

too big to fit into the memory (slide 6). To address this situation, we will need to

divide matrix M into vertical stripes, and also divide v into equal number of

horizontal stripes:

Your task is to design a mapReduce algorithm and implement the algorithm to

divide matrix M into vertical stripes.

Again, matrix M is stored in a file, using (i,j,mij) format. Here is one example matrix,

1 7

11 12 16

22 23 28

30 34

42 45

51 56

63 67

71 78

80 82 84 86 88

and the actual file that stores this matrix is given by the following (you can find this

file from our course website. )

(0,1,1)

(8,0,80)

(0,7,7)

(1,1,11)

(3,0,30)

(2,2,22)

(8,4,84)

(2,8,28)

(3,4,34)

(6,3,63)

(4,5,45)

(5,1,51)

(8,8,88)

(5,6,56)

(6,7,67)

(7,1,71)

(1,6,16)

(7,8,78)

(1,2,12)

(8,2,82)

(2,3,23)

(8,6,86)

(4,2,42)

As you can tell, this matrix is 9 x 9, and we would like to break it into 3 vertical

stripes (therefore each stripe is 9 x 3). The final result will be 3 separate files, each

file holds one stripe from the original matrix.

You will need to submit the following:

a very brief description of your design. The description should be in this

format (20 points)

map task: for each , emits ,

reduce task:

your source code (30 points)

a screen shot showing the command you use to start your job (5 points)

a screen shot showing the result of your submitted job, should be something

look like this (5 points)

hint: you will need to write a customer partitioner, something like this:

public static class MatrixConvertorPartitioner extends

Partitioner