Lambda architecture (batch + streaming) in Hadoop

Introduction

Lambda architecture helps process large data by combining fast real-time data and slower batch data. It gives quick and accurate results.

When you want to analyze data as it arrives and also keep a full history.

When you need quick insights but also want to correct errors later.

When your data is too big to process all at once.

When you want to combine live updates with detailed reports.

When you want a system that is fault-tolerant and scalable.

Syntax

Hadoop

Lambda Architecture has three layers:

1. Batch Layer:
   - Stores all raw data.
   - Processes data in large batches.
   - Creates a master dataset and batch views.

2. Speed Layer (Streaming Layer):
   - Processes data in real-time.
   - Handles recent data quickly.
   - Creates real-time views.

3. Serving Layer:
   - Merges batch views and real-time views.
   - Answers queries combining both data sources.

The batch layer handles big data but is slower.

The speed layer handles new data fast but with less accuracy.

Examples

This example shows how Hadoop tools fit into Lambda architecture.

Hadoop

Batch Layer: Use Hadoop MapReduce to process all data every hour.
Speed Layer: Use Apache Storm to process data as it arrives.
Serving Layer: Use HBase to combine batch and speed views for queries.

This example uses Kafka and Spark Streaming for real-time data.

Hadoop

Batch Layer: Store raw logs in HDFS.
Speed Layer: Stream logs with Kafka and process with Spark Streaming.
Serving Layer: Query combined results with Apache Hive.

Sample Program

This code shows a simple Lambda architecture example using Spark on Hadoop. Batch data is processed once, streaming data is processed continuously. Results are printed to console.

Hadoop

from pyspark.sql import SparkSession
from pyspark.sql.functions import window, col

# Initialize Spark session
spark = SparkSession.builder.appName('LambdaExample').getOrCreate()

# Batch Layer: Read historical data from HDFS
batch_data = spark.read.json('hdfs://path/to/historical/data')

# Process batch data: count events per day
batch_counts = batch_data.groupBy('eventType').count()

# Speed Layer: Simulate streaming data
streaming_data = spark.readStream.schema(batch_data.schema).json('hdfs://path/to/streaming/data')

# Process streaming data: count events in 1 minute windows
stream_counts = streaming_data.groupBy(window(col('timestamp'), '1 minute'), 'eventType').count()

# Start streaming query to console (for demo)
query = stream_counts.writeStream.outputMode('complete').format('console').start()

# Show batch results
batch_counts.show()

# Wait for streaming to finish (in real use, streaming runs continuously)
query.awaitTermination(10)

spark.stop()

Important Notes

Batch layer is slow but accurate.

Speed layer is fast but may have approximate data.

Serving layer combines both for best results.

Summary

Lambda architecture mixes batch and streaming to handle big data.

Batch layer processes all data; speed layer processes new data fast.

Serving layer merges results for queries.

Practice

1. What is the main purpose of the batch layer in Lambda architecture?

easy

A. To process all data for accurate and complete results

B. To handle real-time data streams quickly

C. To store raw data without processing

D. To visualize data for end users

Solution

1. Step 1: Understand batch layer role

   The batch layer processes all accumulated data to provide accurate and comprehensive views.

2. Step 2: Compare with other layers

   The speed layer handles real-time data, not the batch layer.

3. Final Answer:

   To process all data for accurate and complete results -> Option A

4. Quick Check:

   Batch layer = accurate full data processing [OK]

Hint: Batch layer = full data, accurate results [OK]

Common Mistakes:

• Confusing batch layer with speed layer
• Thinking batch layer processes data in real-time
• Assuming batch layer only stores raw data

2. Which of the following is the correct component of Lambda architecture for handling real-time data?

easy

A. Batch layer

B. Speed layer

C. Serving layer

D. Storage layer

Solution

1. Step 1: Identify real-time processing component

   The speed layer is designed to process data streams in real-time for quick insights.

2. Step 2: Eliminate other options

   Batch layer processes full data, serving layer provides query results, storage layer is not a standard Lambda component.

3. Final Answer:

   Speed layer -> Option B

4. Quick Check:

   Speed layer = real-time data processing [OK]

Hint: Speed layer = real-time data [OK]

Common Mistakes:

• Choosing batch layer for real-time data
• Confusing serving layer with speed layer
• Thinking storage layer is a Lambda component

3. Consider this simplified pseudocode for Lambda architecture layers:

batch_result = batch_layer(data)
speed_result = speed_layer(stream)
final_result = merge(batch_result, speed_result)
print(final_result)

If batch_layer returns {'count': 100} and speed_layer returns {'count': 5}, what will final_result be after merging by summing counts?

medium

A. {'count': 105}

B. {'count': 100}

C. {'count': 5}

D. Error: Cannot merge dictionaries

Solution

1. Step 1: Understand merging logic

   Merging sums the 'count' values from batch and speed results: 100 + 5 = 105.

2. Step 2: Confirm final result

   The final dictionary after merging is {'count': 105}.

3. Final Answer:

   {'count': 105} -> Option A

4. Quick Check:

   100 + 5 = 105 [OK]

Hint: Sum batch and speed counts for final result [OK]

Common Mistakes:

• Ignoring speed layer result
• Assuming merge replaces batch with speed
• Thinking merge causes error

4. This code snippet tries to implement the speed layer but has an error:

def speed_layer(stream):
    for record in stream:
        process(record)
    return processed_data

stream = get_stream()
result = speed_layer(stream)
print(result)

What is the likely error here?

medium

A. process() function is missing

B. stream is not iterable

C. processed_data is not defined before return

D. print() syntax error

Solution

1. Step 1: Check variable initialization

   The variable processed_data is returned but never defined or updated inside the function.

2. Step 2: Verify other parts

   Assuming stream is iterable and process() exists, the main error is missing processed_data definition.

3. Final Answer:

   processed_data is not defined before return -> Option C

4. Quick Check:

   Undefined variable causes error [OK]

Hint: Check all variables are defined before return [OK]

Common Mistakes:

• Assuming process() is missing without checking
• Thinking stream is not iterable
• Ignoring variable initialization

5. In a Lambda architecture, if the batch layer updates data every 24 hours and the speed layer processes data every second, what is the best way to ensure the serving layer shows the most accurate and up-to-date results?

hard

A. Use the speed layer results immediately and replace them with batch results after 24 hours

B. Only show batch layer results to avoid inconsistency

C. Ignore speed layer and rely on batch layer for all queries

D. Merge speed and batch results continuously, updating serving layer with combined data

Solution

1. Step 1: Understand serving layer role

   The serving layer combines batch and speed layer outputs to provide accurate and timely results.

2. Step 2: Evaluate options

   Merge speed and batch results continuously, updating serving layer with combined data merges both results continuously, ensuring the serving layer is both up-to-date and accurate.

3. Final Answer:

   Merge speed and batch results continuously, updating serving layer with combined data -> Option D

4. Quick Check:

   Serving layer = combined batch + speed data [OK]

Hint: Serving layer merges batch and speed data continuously [OK]

Common Mistakes:

• Using only speed or batch layer results alone
• Replacing speed results abruptly without merging
• Ignoring serving layer's combining role