Why distributed databases handle scale in DBMS Theory - The Real Reasons

The Big Idea

What if your favorite app could never slow down, no matter how many people use it at once?

The Scenario

Imagine a small shop trying to keep track of all its sales and customers using just one notebook. As the shop grows into a big mall with thousands of customers, that single notebook becomes too small and slow to handle all the information.

The Problem

Using one notebook (or one database server) means everything must be written and read from the same place. This slows down the process, causes delays, and if the notebook is lost or damaged, all data is lost. It also can't handle many people using it at once without mistakes or crashes.

The Solution

Distributed databases split the data across many computers working together. This way, they share the load, work faster, and keep data safe even if one computer fails. It's like having many notebooks, each handling part of the work, making the whole system stronger and quicker.

Before vs After

✗ Before

SELECT * FROM sales WHERE date = '2024-01-01'; -- single server handles all queries

✓ After

SELECT * FROM sales WHERE date = '2024-01-01'; -- query distributed across multiple servers

What It Enables

Distributed databases enable systems to grow smoothly and handle huge amounts of data and users without slowing down or breaking.

Real Life Example

Big online stores like Amazon use distributed databases to manage millions of products and customers worldwide, ensuring fast and reliable shopping experiences.

Key Takeaways

Single databases struggle with large data and many users.

Distributed databases spread data and work across many machines.

This approach improves speed, reliability, and scalability.

Practice

1. Why do distributed databases handle scale better than single-server databases?

easy

A. Because they spread data and workload across multiple machines

B. Because they use only one powerful computer

C. Because they store data in a single location

D. Because they limit the number of users accessing data

Solution

1. Step 1: Understand the concept of distributed databases

   Distributed databases store data on many computers instead of just one.

2. Step 2: Recognize how spreading data helps scale

   Spreading data and workload means many machines share the work, so the system can handle more data and users.

3. Final Answer:

   Because they spread data and workload across multiple machines -> Option A

4. Quick Check:

   Distributed databases = spread data/workload = better scale [OK]

Hint: Think: More machines share work, so system handles more [OK]

Common Mistakes:

• Thinking a single powerful computer is enough
• Believing data stored in one place scales well
• Assuming limiting users improves scaling

2. Which of the following is a correct reason why distributed databases improve reliability?

easy

A. They store all data on a single server

B. They replicate data across multiple nodes

C. They delete old data regularly

D. They restrict access to one user at a time

Solution

1. Step 1: Identify how reliability is improved in distributed systems

   Reliability means data is safe and accessible even if one machine fails.

2. Step 2: Understand data replication

   Replicating data means copying it to multiple machines, so if one fails, others still have the data.

3. Final Answer:

   They replicate data across multiple nodes -> Option B

4. Quick Check:

   Replication = data copies = better reliability [OK]

Hint: Replication means copies on many machines, so safer data [OK]

Common Mistakes:

• Thinking storing data on one server improves reliability
• Confusing deleting data with reliability
• Believing restricting users improves reliability

3. Consider a distributed database system with 4 nodes. If each node can handle 1000 queries per second, what is the total query capacity of the system?

medium

A. 250 queries per second

B. 1000 queries per second

C. 4000 queries per second

D. 5000 queries per second

Solution

1. Step 1: Understand capacity per node

   Each node can handle 1000 queries per second.

2. Step 2: Calculate total capacity by adding all nodes

   4 nodes x 1000 queries = 4000 queries per second total capacity.

3. Final Answer:

   4000 queries per second -> Option C

4. Quick Check:

   4 x 1000 = 4000 queries/sec [OK]

Hint: Multiply nodes by capacity per node for total [OK]

Common Mistakes:

• Using capacity of one node as total
• Dividing instead of multiplying
• Adding extra queries beyond node capacity

4. A distributed database is not scaling well. Which of the following is a likely cause?

medium

A. The database uses multiple machines

B. Data is replicated on all nodes

C. There are too many nodes handling queries

D. Data is not evenly distributed across nodes

Solution

1. Step 1: Identify what causes poor scaling

   Poor scaling happens if some nodes have too much data or work, causing bottlenecks.

2. Step 2: Understand uneven data distribution

   If data is not spread evenly, some nodes get overloaded while others are idle, hurting performance.

3. Final Answer:

   Data is not evenly distributed across nodes -> Option D

4. Quick Check:

   Uneven data = overloaded nodes = poor scaling [OK]

Hint: Check if data is balanced across nodes for good scale [OK]

Common Mistakes:

• Thinking more nodes always cause poor scaling
• Believing replication causes poor scaling
• Assuming multiple machines hurt scaling

5. A company wants to handle a sudden increase in users without slowing down their database. Which distributed database feature should they focus on to handle this scale?

hard

A. Adding more nodes to share the workload

B. Reducing data replication to save space

C. Storing all data on a single powerful server

D. Limiting user access during peak times

Solution

1. Step 1: Understand the need to handle more users

   More users mean more queries and data requests, requiring more processing power.

2. Step 2: Identify how distributed databases handle increased load

   Adding more nodes spreads the workload, so the system can handle more users without slowing down.

3. Final Answer:

   Adding more nodes to share the workload -> Option A

4. Quick Check:

   More nodes = shared workload = better scaling [OK]

Hint: Add nodes to share work and handle more users [OK]

Common Mistakes:

• Thinking reducing replication improves scaling
• Believing one powerful server can handle all load
• Assuming limiting users is the best scaling method