Read replicas in GCP - Time & Space Complexity

Time Complexity: Read replicas

O(n)

Understanding Time Complexity

When using read replicas in cloud databases, it's important to understand how the number of replicas affects the work done by the system.

We want to know how the system's operations grow as we add more read replicas.

Scenario Under Consideration

Analyze the time complexity of the following operation sequence.

// Create a primary database instance
const primaryInstance = createInstance('primary-db');

// Create multiple read replicas
for (let i = 0; i < n; i++) {
  createReadReplica(primaryInstance, `replica-${i}`);
}

// Read queries are sent to replicas
sendReadQuery(`replica-${queryTarget}`);

This sequence creates one primary database and then creates n read replicas from it. Read queries are distributed to these replicas.

Identify Repeating Operations

Identify the API calls, resource provisioning, data transfers that repeat.

• Primary operation: Creating each read replica instance.
• How many times: Once for each replica, so n times.
• Data replication: Continuous data syncing from primary to each replica, happening n times in parallel.

How Execution Grows With Input

As you add more replicas, the system must create and maintain each one separately.

Input Size (n)	Approx. Api Calls/Operations
10	About 10 replica creation calls and 10 data sync streams
100	About 100 replica creation calls and 100 data sync streams
1000	About 1000 replica creation calls and 1000 data sync streams

Pattern observation: The work grows directly with the number of replicas added.

Final Time Complexity

Time Complexity: O(n)

This means the time and work needed grow in a straight line as you add more read replicas.

Common Mistake

[X] Wrong: "Adding more read replicas does not increase system work because they just copy data automatically."

[OK] Correct: Each replica requires separate setup and continuous syncing, so the system's work increases with each one.

Interview Connect

Understanding how adding replicas affects system work helps you design scalable cloud databases and explain trade-offs clearly.

Self-Check

"What if we used a caching layer instead of read replicas? How would the time complexity of handling read requests change?"

Practice

1.

What is the main purpose of a read replica in Google Cloud SQL?

easy

A. To copy the main database and handle read requests to reduce load

B. To store backups of the database for disaster recovery

C. To serve as a primary database for write operations

D. To encrypt data stored in the database

Solution

1. Step 1: Understand the role of read replicas

   Read replicas copy data from the main database and serve read requests to reduce load on the primary instance.

2. Step 2: Differentiate from other database roles

   Backups and encryption are different features; read replicas do not handle writes.

3. Final Answer:

   To copy the main database and handle read requests to reduce load -> Option A

4. Quick Check:

   Read replica purpose = reduce load by handling reads [OK]

Hint: Read replicas handle reads to reduce main database load [OK]

Common Mistakes:

• Confusing read replicas with backups
• Thinking read replicas handle writes
• Assuming read replicas encrypt data

2.

Which gcloud command correctly creates a read replica named replica-instance of a master instance called master-instance?

easy

A. gcloud sql instances create replica-instance --master-instance-name=master-instance

B. gcloud sql instances create replica-instance --replica-of=master-instance

C. gcloud sql replicas create replica-instance --master=master-instance

D. gcloud sql create instance replica-instance --master=master-instance

Solution

1. Step 1: Recall the correct command syntax

   The correct command to create a read replica uses gcloud sql instances create with the flag --master-instance-name.

2. Step 2: Verify option correctness

   gcloud sql instances create replica-instance --master-instance-name=master-instance matches the correct syntax exactly; other options use incorrect flags or commands.

3. Final Answer:

   gcloud sql instances create replica-instance --master-instance-name=master-instance -> Option A

4. Quick Check:

   Correct flag for read replica creation = --master-instance-name [OK]

Hint: Use --master-instance-name flag with gcloud sql instances create [OK]

Common Mistakes:

• Using incorrect flags like --replica-of
• Using wrong gcloud subcommands
• Mixing up command order or syntax

3.

Consider this command output snippet after creating a read replica:

NAME              REGION       MASTER_INSTANCE
replica-instance  us-central1  master-instance

What does this output indicate?

medium

A. The master-instance is a read replica of replica-instance

B. The replica-instance is the primary database in us-central1 region

C. The replica-instance is a read replica of master-instance in us-central1 region

D. The replica-instance is a backup stored in us-central1

Solution

1. Step 1: Analyze the output columns

   The output shows the replica name, its region, and the master instance it replicates.

2. Step 2: Interpret the relationship

   Since replica-instance lists master-instance as its master, it is a read replica of that master in the specified region.

3. Final Answer:

   The replica-instance is a read replica of master-instance in us-central1 region -> Option C

4. Quick Check:

   Replica points to master-instance = read replica relationship [OK]

Hint: Master instance listed means replica points to it for reads [OK]

Common Mistakes:

• Confusing replica as primary
• Thinking master is replica
• Assuming backup instead of replica

4.

You tried to create a read replica with this command:

gcloud sql instances create replica1 --master=master1

But it failed. What is the likely cause?

medium

A. You must create the replica before the master instance

B. The replica name cannot contain numbers

C. The master instance must be in a different region

D. The flag --master is incorrect; should use --master-instance-name

Solution

1. Step 1: Check the command flags

   The correct flag to specify the master instance is --master-instance-name, not --master.

2. Step 2: Validate other options

   Replica names can contain numbers; master and replica can be in same region; master must exist before replica.

3. Final Answer:

   The flag --master is incorrect; should use --master-instance-name -> Option D

4. Quick Check:

   Correct flag for master instance = --master-instance-name [OK]

Hint: Use exact flag --master-instance-name to link replica to master [OK]

Common Mistakes:

• Using wrong flag --master
• Thinking replica name rules are stricter
• Assuming region mismatch causes failure

5.

You have a master instance handling both reads and writes. To improve read performance globally, you want to create two read replicas in different regions. Which approach is best?

1. Create two read replicas using gcloud sql instances create with --master-instance-name for each region.
2. Configure your application to send read queries to replicas and write queries to the master.
3. Ensure replicas are in regions close to your users for low latency.

What is the correct reasoning behind this approach?

hard

A. Read replicas automatically handle writes to reduce master load

B. Read replicas reduce load on master and improve read latency by being regionally closer to users

C. Creating replicas in multiple regions duplicates writes to all replicas instantly

D. Replicas replace the master for all database operations

Solution

1. Step 1: Understand read replica benefits

   Read replicas copy data from the master and serve read requests, reducing load on the master.

2. Step 2: Consider regional placement and query routing

   Placing replicas near users lowers latency; routing reads to replicas and writes to master optimizes performance.

3. Step 3: Clarify write behavior

   Replicas do not handle writes; writes go only to the master and replicate asynchronously.

4. Final Answer:

   Read replicas reduce load on master and improve read latency by being regionally closer to users -> Option B

5. Quick Check:

   Replicas improve read speed and reduce master load [OK]

Hint: Replicas serve reads near users; writes stay on master [OK]

Common Mistakes:

• Assuming replicas handle writes
• Thinking replicas instantly sync writes synchronously
• Believing replicas replace the master