Kafka vs RabbitMQ vs Redis Pub/Sub - Performance Comparison

Time Complexity: Kafka vs RabbitMQ vs Redis Pub/Sub

O(n * p)

Understanding Time Complexity

When working with messaging systems like Kafka, RabbitMQ, and Redis Pub/Sub, it's important to understand how their processing time grows as the number of messages or clients increases.

We want to know how the time to send and receive messages changes when the workload grows.

Scenario Under Consideration

Analyze the time complexity of sending messages to multiple consumers using Kafka.

    // Pseudocode for Kafka message publishing
    for each message in messages:
        for each partition in topic_partitions:
            send message to partition
    

This code sends each message to all partitions of a Kafka topic, simulating message distribution.

Identify Repeating Operations

Look at the loops that repeat work:

• Primary operation: Sending messages to partitions.
• How many times: For each message, it sends to every partition.

How Execution Grows With Input

As the number of messages (n) or partitions (p) grows, the total sends grow roughly by n times p.

Input Size (messages n)	Approx. Operations (n * p)
10 messages, 3 partitions	30 sends
100 messages, 3 partitions	300 sends
1000 messages, 3 partitions	3000 sends

Pattern observation: The work grows linearly with both messages and partitions.

Final Time Complexity

Time Complexity: O(n * p)

This means the time to send messages grows proportionally with the number of messages and the number of partitions.

Common Mistake

[X] Wrong: "Sending messages to multiple consumers is always constant time regardless of consumers."

[OK] Correct: Each consumer or partition usually requires separate processing, so time grows with how many there are.

Interview Connect

Understanding how message systems scale helps you design efficient data flows and troubleshoot delays in real projects.

Self-Check

What if we changed from sending messages to all partitions to sending only to one partition? How would the time complexity change?

Practice

1. Which messaging system is best suited for storing message history and handling large data streams?

easy

A. None of the above

B. RabbitMQ

C. Redis Pub/Sub

D. Kafka

Solution

1. Step 1: Understand Kafka's purpose

   Kafka is designed to handle large data streams and keeps a durable log of messages, allowing message history to be stored.

2. Step 2: Compare with RabbitMQ and Redis Pub/Sub

   RabbitMQ focuses on reliable delivery and routing, while Redis Pub/Sub is fast but does not store messages.

3. Final Answer:

   Kafka -> Option D

4. Quick Check:

   Message history and big data = Kafka [OK]

Hint: Remember: Kafka stores history; Redis Pub/Sub does not [OK]

Common Mistakes:

• Confusing Redis Pub/Sub as storing messages
• Thinking RabbitMQ stores message history
• Assuming all systems keep message logs

2. Which of the following is the correct statement about Redis Pub/Sub?

easy

A. Redis Pub/Sub guarantees message delivery even if subscribers are offline.

B. Redis Pub/Sub does not store messages; it only delivers messages to active subscribers.

C. Redis Pub/Sub supports complex message routing like RabbitMQ.

D. Redis Pub/Sub is designed for batch processing of messages.

Solution

1. Step 1: Analyze Redis Pub/Sub behavior

   Redis Pub/Sub delivers messages only to subscribers currently connected; it does not store messages for offline clients.

2. Step 2: Check other options

   Redis Pub/Sub guarantees message delivery even if subscribers are offline. is false because Redis Pub/Sub does not guarantee delivery if subscribers are offline. Redis Pub/Sub supports complex message routing like RabbitMQ. is false because Redis Pub/Sub lacks complex routing. Redis Pub/Sub is designed for batch processing of messages. is false because Redis Pub/Sub is for real-time messaging, not batch processing.

3. Final Answer:

   Redis Pub/Sub does not store messages; it only delivers messages to active subscribers. -> Option B

4. Quick Check:

   Redis Pub/Sub delivers only to active subscribers [OK]

Hint: Redis Pub/Sub = no message storage, only live delivery [OK]

Common Mistakes:

• Assuming Redis Pub/Sub stores messages
• Confusing Redis with RabbitMQ routing features
• Thinking Redis Pub/Sub supports offline message delivery

3. Given the following scenario: You need a messaging system that supports flexible routing and guarantees message delivery even if consumers are temporarily offline. Which system fits best?

medium

A. Kafka

B. Redis Pub/Sub

C. RabbitMQ

D. None of the above

Solution

1. Step 1: Understand RabbitMQ features

   RabbitMQ supports flexible routing through exchanges and queues and guarantees message delivery with acknowledgments and persistence.

2. Step 2: Compare with Kafka and Redis Pub/Sub

   Kafka stores messages but is less flexible in routing. Redis Pub/Sub is fast but does not guarantee delivery if consumers are offline.

3. Final Answer:

   RabbitMQ -> Option C

4. Quick Check:

   Flexible routing + guaranteed delivery = RabbitMQ [OK]

Hint: Flexible routing and delivery guarantees point to RabbitMQ [OK]

Common Mistakes:

• Choosing Kafka for flexible routing
• Assuming Redis Pub/Sub guarantees delivery
• Ignoring RabbitMQ's message acknowledgment feature

4. You have a system using Redis Pub/Sub, but messages are lost when subscribers disconnect. What is the best way to fix this issue?

medium

A. Switch to Kafka to store messages and allow offline consumption.

B. Increase Redis server memory to store more messages.

C. Use Redis Pub/Sub with message persistence enabled.

D. Configure Redis to retry sending messages to offline subscribers.

Solution

1. Step 1: Identify Redis Pub/Sub limitation

   Redis Pub/Sub does not store messages; if subscribers disconnect, messages are lost.

2. Step 2: Choose a system that stores messages

   Kafka stores messages durably, allowing consumers to read messages even after being offline.

3. Final Answer:

   Switch to Kafka to store messages and allow offline consumption. -> Option A

4. Quick Check:

   Redis Pub/Sub no storage; Kafka stores messages [OK]

Hint: Redis Pub/Sub can't store messages; Kafka can [OK]

Common Mistakes:

• Trying to enable persistence on Redis Pub/Sub (not supported)
• Assuming Redis memory increase solves message loss
• Expecting Redis to retry offline delivery automatically

5. You want to build a real-time analytics system that processes large volumes of data and keeps a full history of all events for auditing. Which combination of messaging systems is the best approach?

hard

A. Use Kafka for event storage and Redis Pub/Sub for real-time notifications.

B. Use RabbitMQ for event storage and Redis Pub/Sub for message routing.

C. Use Redis Pub/Sub alone for both storage and real-time processing.

D. Use RabbitMQ alone for storing large data streams and real-time analytics.

Solution

1. Step 1: Identify requirements

   The system needs to handle large data volumes, keep full event history, and provide real-time notifications.

2. Step 2: Match systems to requirements

   Kafka is ideal for storing large data streams and keeping history. Redis Pub/Sub is fast for real-time notifications but does not store data.

3. Step 3: Evaluate other options

   RabbitMQ is good for routing and delivery but less suited for big data storage. Redis Pub/Sub alone cannot store messages. RabbitMQ alone is not optimized for large data streams.

4. Final Answer:

   Use Kafka for event storage and Redis Pub/Sub for real-time notifications. -> Option A

5. Quick Check:

   Kafka stores history + Redis Pub/Sub fast notifications [OK]

Hint: Combine Kafka storage with Redis Pub/Sub speed [OK]

Common Mistakes:

• Using Redis Pub/Sub alone for storage
• Assuming RabbitMQ handles big data storage well
• Ignoring Kafka's strength in event history