Performance optimization in GCP - Time & Space Complexity

Time Complexity: Performance optimization

O(n)

Understanding Time Complexity

When we optimize performance in cloud systems, we want to know how the work grows as we add more tasks or users.

We ask: How does the time to complete tasks change when the workload increases?

Scenario Under Consideration

Analyze the time complexity of scaling a cloud function to handle multiple requests.

// Pseudocode for handling requests in GCP Cloud Functions
function handleRequests(requests) {
  for (let i = 0; i < requests.length; i++) {
    processRequest(requests[i]);
  }
}

function processRequest(request) {
  // Calls external API
  callExternalAPI(request.data);
}
    

This sequence processes each request by calling an external API once per request.

Identify Repeating Operations

Look at what repeats as the number of requests grows.

• Primary operation: Calling the external API for each request.
• How many times: Once per request, so as many times as there are requests.

How Execution Grows With Input

Each new request adds one more API call, so the total work grows steadily with the number of requests.

Input Size (n)	Approx. Api Calls/Operations
10	10 API calls
100	100 API calls
1000	1000 API calls

Pattern observation: The work grows directly in proportion to the number of requests.

Final Time Complexity

Time Complexity: O(n)

This means the time to complete all requests grows in a straight line as the number of requests increases.

Common Mistake

[X] Wrong: "Adding more requests won't affect total time much because calls happen fast."

[OK] Correct: Each request adds a separate call, so total time adds up and grows with the number of requests.

Interview Connect

Understanding how work grows with input helps you explain system behavior clearly and shows you can think about scaling in real projects.

Self-Check

"What if we batch multiple requests into one API call? How would the time complexity change?"

Practice

1. Which of the following is the best way to improve the performance of a Google Cloud Compute Engine instance?

easy

A. Choose a machine type with more CPUs and memory

B. Use a smaller disk size to reduce cost

C. Disable autoscaling to keep resources fixed

D. Use only standard persistent disks for all workloads

Solution

1. Step 1: Understand machine types impact

   More CPUs and memory allow the instance to handle more work faster.

2. Step 2: Evaluate other options

   Smaller disks reduce storage but do not improve performance; disabling autoscaling limits flexibility; standard disks are slower than SSDs.

3. Final Answer:

   Choose a machine type with more CPUs and memory -> Option A

4. Quick Check:

   More CPU and memory = better performance [OK]

Hint: More CPUs and memory usually mean better performance [OK]

Common Mistakes:

• Thinking smaller disks improve speed
• Disabling autoscaling helps performance
• Ignoring disk type impact

2. Which of the following is the correct way to enable autoscaling on a Google Kubernetes Engine (GKE) cluster?

easy

A. Use the kubectl autoscale command with CPU utilization target

B. Disable autoscaling to save costs

C. Set the number of nodes manually and never change it

D. Use only preemptible VMs without autoscaling

Solution

1. Step 1: Understand autoscaling in GKE

   Autoscaling adjusts the number of pods based on CPU or other metrics automatically.

2. Step 2: Identify correct command

   The kubectl autoscale command sets autoscaling with CPU target, enabling dynamic scaling.

3. Final Answer:

   Use the kubectl autoscale command with CPU utilization target -> Option A

4. Quick Check:

   Autoscale command enables dynamic scaling [OK]

Hint: Use kubectl autoscale with CPU target to enable autoscaling [OK]

Common Mistakes:

• Manually setting fixed node counts
• Disabling autoscaling to save costs
• Using preemptible VMs without autoscaling

3. You have a Compute Engine instance with an attached standard persistent disk. You want to improve disk performance. Which change will have the most impact?

medium

A. Switch to a larger standard persistent disk

B. Switch to a balanced persistent disk

C. Switch to an SSD persistent disk

D. Add more CPUs to the instance

Solution

1. Step 1: Understand disk types and performance

   Standard persistent disks are slower; SSD persistent disks offer much better IOPS and throughput.

2. Step 2: Evaluate options

   Larger standard disks increase capacity but not speed; balanced disks are better than standard but slower than SSD; adding CPUs does not improve disk speed.

3. Final Answer:

   Switch to an SSD persistent disk -> Option C

4. Quick Check:

   SSD disks improve disk performance most [OK]

Hint: SSD disks give best disk speed improvement [OK]

Common Mistakes:

• Confusing disk size with speed
• Thinking more CPUs improve disk speed
• Choosing balanced disks over SSD for max speed

4. A Cloud Function is running slower than expected. You find the function has a memory limit of 256 MB. What is the best way to improve its performance?

medium

A. Reduce the memory allocation to 128 MB

B. Increase the memory allocation to 1 GB

C. Disable logging to speed up execution

D. Change the runtime to an older version

Solution

1. Step 1: Understand memory impact on Cloud Functions

   More memory increases CPU allocation, improving performance.

2. Step 2: Evaluate other options

   Reducing memory slows function; disabling logging has minimal effect; older runtimes may be slower or unsupported.

3. Final Answer:

   Increase the memory allocation to 1 GB -> Option B

4. Quick Check:

   More memory means more CPU and faster execution [OK]

Hint: More memory means more CPU for Cloud Functions [OK]

Common Mistakes:

• Reducing memory to save cost
• Disabling logging expecting big speedup
• Using older runtimes for speed

5. You manage a web app on Google App Engine Standard environment. The app experiences high traffic spikes causing slow responses. Which combination of actions will best optimize performance and cost?

hard

A. Disable Cloud CDN and increase instance memory

B. Set a fixed number of instances and disable autoscaling

C. Use only manual scaling with large instance types

D. Enable autoscaling with a minimum number of instances and use Cloud CDN

Solution

1. Step 1: Understand autoscaling benefits

   Autoscaling adjusts instances to traffic, handling spikes efficiently and saving cost during low traffic.

2. Step 2: Use Cloud CDN to reduce backend load

   Cloud CDN caches content closer to users, reducing latency and backend requests.

3. Final Answer:

   Enable autoscaling with a minimum number of instances and use Cloud CDN -> Option D

4. Quick Check:

   Autoscaling + CDN = best performance and cost [OK]

Hint: Combine autoscaling and CDN for best speed and cost [OK]

Common Mistakes:

• Fixing instance count causes wasted resources
• Manual scaling ignores traffic spikes
• Disabling CDN increases backend load