Building Your First Big Data Application on AWS

Matt Yanchyshyn, Sr. Manager, Solutions Architecture

April 2016

2016, Amazon Web Services, Inc. or its Affiliates. All rights reserved.

Your First Big Data Application on AWS

COLLECT
STORE

PROCESS

ANALYZE & VISUALIZE

Your First Big Data Application on AWS

COLLECT:
Amazon Kinesis Firehose

STORE

PROCESS:
Amazon EMR with Spark & Hive

ANALYZE & VISUALIZE:

Amazon Redshift and Amazon QuickSight

A Modern Take on the Classic Data Warehouse

http://aws.amazon.com/big-data/use-cases/

Setting up the environment

Data Storage with Amazon S3

Download all the CLI steps: http://bit.ly/aws-big-data-steps
Create an Amazon S3 bucket to store the data collected
with Amazon Kinesis Firehose
aws s3 mb s3://YOUR-S3-BUCKET-NAME

Access Control with IAM

Create an IAM role to allow Firehose to write to the S3 bucket
firehose-policy.json:
{
"Version": "2012-10-17",
"Statement": {
"Effect": "Allow",
"Principal": {"Service": "firehose.amazonaws.com"},
"Action": "sts:AssumeRole"
}
}

Access Control with IAM

Create an IAM role to allow Firehose to write to the S3 bucket
s3-rw-policy.json:
{ "Version": "2012-10-17",
"Statement": {
"Effect": "Allow",
"Action": "s3:*",
"Resource": [
"arn:aws:s3:::YOUR-S3-BUCKET-NAME",
"arn:aws:s3:::YOUR-S3-BUCKET-NAME/*"
]
} }

Access Control with IAM

Create an IAM role to allow Firehose to write to the S3 bucket
aws iam create-role --role-name firehose-demo \
--assume-role-policy-document file://firehose-policy.json

Copy the value in Arn in the output,

e.g., arn:aws:iam::123456789:role/firehose-demo
aws iam put-role-policy --role-name firehose-demo \
--policy-name firehose-s3-rw \
--policy-document file://s3-rw-policy.json

Data Collection with Amazon Kinesis Firehose

Create a Firehose stream for incoming log data
aws firehose create-delivery-stream \
--delivery-stream-name demo-firehose-stream \
--s3-destination-configuration \
RoleARN=YOUR-FIREHOSE-ARN,\
BucketARN="arn:aws:s3:::YOUR-S3-BUCKET-NAME",\
Prefix=firehose\/,\
BufferingHints={IntervalInSeconds=60},\
CompressionFormat=GZIP

Data Processing with Amazon EMR

Launch an Amazon EMR cluster with Spark and Hive
aws emr create-cluster \
--name "demo" \
--release-label emr-4.5.0 \
--instance-type m3.xlarge \
--instance-count 2 \
--ec2-attributes KeyName=YOUR-AWS-SSH-KEY \
--use-default-roles \
--applications Name=Hive Name=Spark Name=Zeppelin-Sandbox

Record your ClusterId from the output.

Data Analysis with Amazon Redshift

Create a single-node Amazon Redshift data warehouse:
aws redshift create-cluster \
--cluster-identifier demo \
--db-name demo \
--node-type dc1.large \
--cluster-type single-node \
--master-username master \
--master-user-password YOUR-REDSHIFT-PASSWORD \
--publicly-accessible \
--port 8192

Collect

Weblogs Common Log Format (CLF)

75.35.230.210 - - [20/Jul/2009:22:22:42 -0700]
"GET /images/pigtrihawk.jpg HTTP/1.1" 200 29236

"http://www.swivel.com/graphs/show/1163466"
"Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.0.11)
Gecko/2009060215 Firefox/3.0.11 (.NET CLR 3.5.30729)"

Writing into Amazon Kinesis Firehose

Download the demo weblog: http://bit.ly/aws-big-data
Open Python and run the following code to import the log into the stream:
import boto3
iam = boto3.client('iam')
firehose = boto3.client('firehose')
with open('weblog', 'r') as f:
for line in f:
firehose.put_record(
DeliveryStreamName='demo-firehose-stream',
Record={'Data': line}
)
print 'Record added'

Process

Apache Spark
Fast, general purpose engine
for large-scale data processing
Write applications quickly in
Java, Scala, or Python
Combine SQL, streaming, and
complex analytics

Spark SQL
Spark's module for working with structured data using SQL

Run unmodified Hive queries on existing data

Apache Zeppelin

Web-based notebook for interactive analytics

Multiple language backend
Apache Spark integration
Data visualization
Collaboration

https://zeppelin.incubator.apache.org/

View the Output Files in Amazon S3

After about 1 minute, you should see files in your S3
bucket:
aws s3 ls s3://YOUR-S3-BUCKET-NAME/firehose/ --recursive

Connect to Your EMR Cluster and Zeppelin

aws emr describe-cluster --cluster-id YOUR-EMR-CLUSTER-ID

Copy the MasterPublicDnsName. Use port forwarding so you can

access Zeppelin at http://localhost:18890 on your local machine.
ssh -i PATH-TO-YOUR-SSH-KEY -L 18890:localhost:8890 \
hadoop@YOUR-EMR-DNS-NAME

Open Zeppelin with your local web browser and create a new Note:
http://localhost:18890

Exploring the Data in Amazon S3 using Spark

Download the Zeppelin notebook: http://bit.ly/aws-big-data-zeppelin
// Load all the files from S3 into a RDD
val accessLogLines = sc.textFile("s3://YOUR-S3-BUCKET-NAME/firehose/*/*/*/*")
// Count the lines
accessLogLines.count

// Print one line as a string

accessLogLines.first
// delimited by space so split them into fields
var accessLogFields = accessLogLines.map(_.split(" ").map(_.trim))
// Print the fields of a line
accessLogFields.first

Combine Fields: A, B, C A B C
var accessLogColumns = accessLogFields
.map( arrayOfFields => { var temp1 =""; for (field <- arrayOfFields) yield {
var temp2 = ""
if (temp1.replaceAll("\\[","\"").startsWith("\"") && !temp1.endsWith("\""))
temp1 = temp1 + " " + field.replaceAll("\\[|\\]","\"")
else temp1 = field.replaceAll("\\[|\\]","\"")
temp2 = temp1
if (temp1.endsWith("\"")) temp1 = ""
temp2
}})
.map( fields => fields.filter(field => (field.startsWith("\"") &&
field.endsWith("\"")) || !field.startsWith("\"") ))
.map(fields => fields.map(_.replaceAll("\"","")))

Create a Data Frame and Transform the Data

import java.sql.Timestamp
import java.net.URL
case class accessLogs(
ipAddress: String,
requestTime: Timestamp,
requestMethod: String,
requestPath: String,
requestProtocol: String,
responseCode: String,
responseSize: String,
referrerHost: String,
userAgent: String
)

Create a Data Frame and Transform the Data

val accessLogsDF = accessLogColumns.map(line => {
var ipAddress
= line(0)
var requestTime
= new Timestamp(new
java.text.SimpleDateFormat("dd/MMM/yyyy:HH:mm:ss Z").parse(line(3)).getTime())
var requestString
= line(4).split(" ").map(_.trim())
var requestMethod
= if (line(4).toString() != "-") requestString(0) else ""
var requestPath
= if (line(4).toString() != "-") requestString(1) else ""
var requestProtocol = if (line(4).toString() != "-") requestString(2) else ""
var responseCode
= line(5).replaceAll("-","")
var responseSize
= line(6).replaceAll("-","")
var referrerHost
= line(7)
var userAgent
= line(8)
accessLogs(ipAddress, requestTime, requestMethod, requestPath,
requestProtocol,responseCode, responseSize, referrerHost, userAgent)
}).toDF()

Create an External Table Backed by Amazon S3

%sql
CREATE EXTERNAL TABLE access_logs
(
ip_address String,
request_time Timestamp,
request_method String,
request_path String,
request_protocol String,
response_code String,
response_size String,
referrer_host String,
user_agent String
)
PARTITIONED BY (year STRING,month STRING, day STRING)
ROW FORMAT DELIMITED FIELDS TERMINATED BY '\t'
STORED AS TEXTFILE
LOCATION 's3://YOUR-S3-BUCKET-NAME/access-log-processed'

Configure Hive Partitioning and Compression

// set up Hive's "dynamic partitioning
%sql
SET hive.exec.dynamic.partition=true

// compress output files on Amazon S3 using Gzip

%sql
SET hive.exec.compress.output=true
%sql
SET mapred.output.compression.codec=org.apache.hadoop.io.compress.GzipCodec
%sql
SET io.compression.codecs=org.apache.hadoop.io.compress.GzipCodec
%sql
SET hive.exec.dynamic.partition.mode=nonstrict;

Write Output to Amazon S3

import org.apache.spark.sql.SaveMode
accessLogsDF
.withColumn("year", year(accessLogsDF("requestTime")))
.withColumn("month", month(accessLogsDF("requestTime")))
.withColumn("day", dayofmonth(accessLogsDF("requestTime")))
.write
.partitionBy("year","month","day")
.mode(SaveMode.Overwrite)
.insertInto("access_logs")

Query the Data Using Spark SQL

// Check the count of records
%sql
select count(*) from access_log_processed

// Fetch the first 10 records

%sql
select * from access_log_processed limit 10

View the Output Files in Amazon S3

Leave Zeppelin and go back to the console
List the partition prefixes and output files:
aws s3 ls s3://YOUR-S3-BUCKET-NAME/access-log-processed/ \
--recursive

Analyze

Connect to Amazon Redshift

Using the PostgreSQL CLI

psql -h YOUR-REDSHIFT-ENDPOINT \
-p 8192 -U master demo
Or use any JDBC or ODBC SQL client with the PostgreSQL 8.x drivers
or native Amazon Redshift support
Aginity Workbench for Amazon Redshift
SQL Workbench/J

Create an Amazon Redshift Table to Hold Your Data

CREATE TABLE accesslogs
(
host_address varchar(512),
request_time timestamp,
request_method varchar(5),
request_path varchar(1024),
request_protocol varchar(10),
response_code Int,
response_size Int,
referrer_host varchar(1024),
user_agent varchar(512)
)
DISTKEY(host_address)
SORTKEY(request_time);

Loading Data into Amazon Redshift

COPY command loads files in parallel from
Amazon S3:
COPY accesslogs
FROM 's3://YOUR-S3-BUCKET-NAME/access-log-processed'
CREDENTIALS
'aws_access_key_id=YOUR-IAMACCESS_KEY;aws_secret_access_key=YOUR-IAM-SECRET-KEY'
DELIMITER '\t'
MAXERROR 0
GZIP;

Amazon Redshift Test Queries

-- find distribution of response codes over days
SELECT TRUNC(request_time), response_code, COUNT(1) FROM
accesslogs GROUP BY 1,2 ORDER BY 1,3 DESC;
-- find the 404 status codes
SELECT COUNT(1) FROM accessLogs WHERE response_code = 404;
-- show all requests for status as PAGE NOT FOUND
SELECT TOP 1 request_path,COUNT(1) FROM accesslogs WHERE
response_code = 404 GROUP BY 1 ORDER BY 2 DESC;

Visualize the Results

DEMO
Amazon QuickSight

Automating the Big Data Application

Amazon
Kinesis
Firehose

Amazon
S3

Event
Notification

Amazon
EMR

Amazon
S3

Amazon
Redshift

Spark job
List of objects from Lambda
Write to Amazon Redshift using spark-redshift

Amazon
QuickSight

Learn from big data experts

blogs.aws.amazon.com/bigdata