CDN Foundations & Core Routing Quiz

Q1. Why does fetching content from a distant origin server typically result in higher latency compared to using a nearby CDN edge server? Because internet protocols add special delays when crossing multiple regional networks.
Because extra DNS lookups are required for distant origin servers, slowing down the response.
Because the data has to travel a longer physical distance over the network, adding propagation delay.
Because the origin server's CPU takes longer to process requests from faraway users.

Q2. Which statement correctly compares Anycast routing with Geo-DNS routing in CDNs? Anycast uses one global IP address advertised from many PoPs so that users are routed (via BGP) to the nearest location, whereas Geo-DNS returns region-specific IPs based on the user's DNS query location.
Using Anycast requires assigning different IP addresses for each geographical region, unlike Geo-DNS.
Anycast provides region-specific DNS responses, while Geo-DNS relies on BGP announcements to route traffic.
Geo-DNS typically fails over faster than Anycast because DNS changes propagate almost instantly on the internet.

Q3. What is the primary benefit of using an origin shield (mid-tier cache) in a CDN's architecture? It ensures content is permanently stored in cache so it never expires or gets evicted.
It reduces load on the origin by having a centralized cache that aggregates cache misses from edge locations, so the origin sees fewer duplicate requests.
It serves content directly to end-users from within the origin's data center network.
It adds an extra encryption layer between the CDN and origin for enhanced security.

Q4. By default, which components of an HTTP request typically make up a CDN cache key? The HTTP method and the content of the request body.
Cookie values and authorization headers from the request.
The client's IP address and the request's User-Agent header.
The request's host (domain) together with the path and query string.

Q5. Which HTTP response header instructs a CDN to cache separate variants of a resource based on certain request attributes (for example, caching different content per `User-Agent` or `Accept-Language`)? ETag
Content-Type
Vary
Cache-Control

Q6. Which HTTP header directive would you use to tell a CDN to cache a response for up to one hour? Cache-Control: no-store
Cache-Control: max-age=3600
ETag: 3600
Cache-Control: no-cache

Q7. What is the purpose of the ETag HTTP header in the context of CDN caching? It tags the content so the CDN knows which objects to purge together.
It indicates which content variant (e.g., language or device type) the CDN should serve.
It tells the CDN how long to cache the resource before it expires.
It serves as a unique identifier for a specific version of a resource, allowing a CDN to validate its cached content with the origin (via If-None-Match) instead of refetching the whole object.

Q8. In a CDN, what is the difference between a hard purge and a soft purge of cached content? Hard purges only affect one geographic region at a time, whereas soft purges apply globally across all CDN PoPs.
A hard purge immediately invalidates and removes the content from cache (so it won't be served until fetched again), whereas a soft purge marks the content as stale but allows it to be served (and revalidated in the background).
A soft purge completely deletes the content from all edge servers, while a hard purge leaves it in cache but hidden for a while.
Soft purges are automatic when content expires, whereas hard purges must be triggered manually.

Q9. What is an advantage of tag-based cache invalidation (purging by tag) in a CDN? It automatically refreshes any content with that tag whenever one item is updated, without needing a purge call.
It significantly increases the cache lifetime (TTL) of the tagged content.
It is the only way to invalidate content before its TTL expires.
It allows you to invalidate a group of related cached items with a single request by assigning them a common tag (so you don't have to purge each URL one by one).

Q10. Why do content delivery networks typically terminate TLS (SSL) connections at the edge servers? Because browsers only trust certificates presented by CDN edge servers.
To reduce latency for end-users by completing the TLS handshake close to the user, thereby avoiding long round trips to the origin for encryption setup.
Because CDNs cannot pass encrypted traffic to the origin servers.
Because terminating TLS at the edge removes the need to encrypt traffic between the edge and the origin.

Q11. Which of the following is a key improvement of the QUIC protocol (used for HTTP/3) compared to traditional HTTP/1.1 or HTTP/2 over TCP? It does not use any encryption, making it faster by avoiding the overhead of TLS.
It runs on top of TCP to ensure reliable delivery, combining the benefits of TCP and UDP.
It establishes connections with lower latency by using UDP with built-in encryption, allowing a secure session to be set up with fewer round-trip handshakes.
It guarantees no packet loss during data transmission, unlike TCP which can lose packets.

Q12. In CDN metrics, what does the cache hit ratio represent? The percentage of requests that are served from the cache (without going to the origin) out of the total requests.
The fraction of the CDN's storage capacity currently being used.
The proportion of a cached file that was delivered before the connection closed.
The average time it takes for the CDN to serve a cached object.

Q13. What is the 'cache miss penalty' in the context of CDN performance? The extra latency incurred when the CDN must fetch content from the origin because it was not in cache (compared to a cache hit).
The reduction in cache hit ratio after a piece of content is purged.
A fee charged by the CDN provider each time the origin is contacted for a cache miss.
The longer TTL assigned to content that the CDN had to retrieve from the origin.

Q14. In monitoring CDN performance, what does 'p99 latency' refer to? The maximum recorded latency out of every 99 requests.
The 99th percentile latency — the time under which 99% of all requests are completed (only 1% of requests take longer than this).
The latency experienced by 99 out of 100 users on the site.
The average latency of the fastest 99% of requests.

Q15. Which statement correctly describes Amazon CloudFront's caching infrastructure? CloudFront edge locations are only present in a few AWS Regions, rather than globally distributed.
Regional Edge Caches are origin servers provided by AWS that you must use for CloudFront distributions.
CloudFront has a global network of edge locations (PoPs) that serve content to users, and it also uses regional edge caches as a second-tier cache between the edge locations and the origin to improve cache efficiency and reduce origin load.
If an edge location cache misses, CloudFront always goes directly to the origin because it does not have any mid-tier caches.

system-design