Lecture on Content Delivery Networks (CDNs)

Introduction

• Problem: Website load time varies by location (e.g., US vs. China).
• Solution: Strategies like CDNs minimize request latency, crucial for global systems design.

What are CDNs?

• Definition: CDNs are networks that distribute and deliver content globally.
• Purpose: Reduce latency by caching static assets closer to users.
• Components:
  • Servers spread globally.
  • Cache static assets like images, videos, HTML, CSS, JavaScript.

Benefits of Using CDNs

• Latency Reduction: Users fetch assets from nearby CDN rather than distant origin servers.
• Bandwidth Savings: Large files (e.g., HD images) use less network bandwidth.

Popular CDNs and Providers

• Cloudflare CDN
• AWS CloudFront
• GCP Cloud CDN
• Azure CDN
• Oracle CDN

How Do CDNs Work?

• Structure: Globally distributed servers with local caches.
• Cache Synchronization: All caches should be up-to-date.

Types of CDNs

1. Push CDNs
   • Engineers push files to CDN.
   • Pros: Ensures assets are up-to-date.
   • Cons: More engineering effort.
2. Pull CDNs
   • Caches updated lazily upon request.
   • Pros: Less maintenance required.
   • Cons:
     • Cache staleness after asset updates.
     • First request latency due to fetching from origin.
   • Popularity: Pull CDNs are more common due to ease of maintenance.

Reducing Cache Staleness

• Timestamping: Cache with expiry times (default often 24 hours).
• Cache Busting: Use unique hashes or E-Tags for asset versions.

When Not to Use CDNs

• Regional Target: No benefit if users are region-specific.
• Dynamic/Sensitive Assets: Avoid stale data (e.g., financial/government services).

Example Use Case

• Interview Question: How would a CDN fit in Amazon’s product listing service?

Conclusion

• CDNs are crucial for reducing latency of static files, but not suited for all scenarios.
• Evaluate the use case (e.g., user region, asset type) before implementing CDNs.