Junior Distributed Caching & CDN Interview Questions (50 MCQs)

Question 1

What is the main goal of adding a cache layer to a system?

Accepted Answer

Reducing latency and load on the source of truth by keeping frequent results closer to the consumer

Answer

Guaranteeing data durability in case the primary database fails

Answer

Fully replacing the primary database with a faster in-memory store

Answer

Guaranteeing strong consistency across all services in a distributed system

Question 2

What does 'cache hit ratio' mean?

Accepted Answer

The percentage of requests served directly from the cache

Answer

The ratio of cache size to total database size

Answer

The number of keys evicted from the cache per second

Answer

The percentage of requests sent to origin that fail

Question 3

What typically happens on a cache miss?

Accepted Answer

Data is read from the source of truth and usually stored in the cache for later requests

Answer

The request is rejected with `404` because the data is not in the cache

Answer

The cache flushes all its contents to make room for the new data

Answer

The client must resend the request directly to the database server

Question 4

Which order lists common cache layers from closest to the user to closest to the data?

Accepted Answer

Browser → CDN → reverse proxy → app cache → distributed cache → DB buffer pool

Answer

CDN → browser → distributed cache → reverse proxy → DB buffer pool → app cache

Answer

Browser → distributed cache → CDN → app cache → reverse proxy → DB buffer pool

Answer

Reverse proxy → browser → CDN → DB buffer pool → app cache → distributed cache

Question 5

What is a CDN (Content Delivery Network)?

Accepted Answer

A geographically distributed network of servers that serves content from the edge nearest the user

Answer

A distributed database that synchronously replicates data across multiple regions

Answer

An internal load balancer that spreads traffic among servers in one datacenter

Answer

A DNS service that only translates domain names to the origin server IP

Question 6

In CDN terminology, what is a 'PoP' (Point of Presence)?

Accepted Answer

A physical location with edge servers that answers requests from users in that area

Answer

The main server where the authoritative copy of the content lives

Answer

A protocol for synchronizing content between edge servers

Answer

An HTTP header that sets how long a response is cached

Question 7

In a CDN architecture, what is the 'origin'?

Accepted Answer

The primary server or storage the CDN fetches content from on a miss

Answer

The edge server nearest the end user that returns the response

Answer

The user's browser that sends the first request for content

Answer

The DNS server that returns the right PoP address to the client

Question 8

Which type of content is the simplest and best fit for CDN caching?

Accepted Answer

Static files such as images, CSS, JS, and fonts

Answer

Personalized API responses such as a user's account balance

Answer

`POST` requests from order submission forms

Answer

WebSocket messages for real-time chat

Question 9

What does a TTL (Time To Live) define in a cache?

Accepted Answer

How long an entry is considered valid before it expires

Answer

The maximum number of times a key can be read from the cache

Answer

How long it takes data to travel from origin to edge

Answer

The time needed to restart a cache node after a crash

Question 10

What trade-off does increasing a cache entry's TTL usually create?

Accepted Answer

Higher hit ratio and less origin load, but more chance of serving stale data

Answer

Fresher data for users, but more CPU usage on cache servers

Answer

Lower write latency, but more risk of losing data in the database

Answer

Better security for sensitive data, but more bandwidth usage

Question 11

What does 'stale' data in a cache mean?

Accepted Answer

Data that is in the cache but no longer matches the current source of truth

Answer

Data that was removed from the cache due to memory pressure

Answer

Data that has never been read from the cache yet

Answer

Data that is stored in compressed form in the cache

Question 12

In the cache-aside (lazy loading) pattern, who is responsible for populating the cache?

Accepted Answer

Application code: on a miss it reads the DB and writes the result to the cache

Answer

The cache itself: on a miss it automatically loads data from the DB

Answer

The database: after every write it pushes changes to the cache

Answer

The CDN: it copies popular content from origin to the internal cache

Question 13

What is the main difference between read-through and cache-aside?

Accepted Answer

In read-through the cache layer itself loads from the source on a miss, not the app code

Answer

In read-through data never expires and has no TTL

Answer

In read-through every write goes to the cache first and then to the DB

Answer

In read-through only static data can be cached

Question 14

How is a write handled in the write-through pattern?

Accepted Answer

Data is written to both cache and DB synchronously before the write is acknowledged

Answer

Data is written only to the cache and later flushed asynchronously to the DB

Answer

Data is written only to the DB and the related cache entry is deleted

Answer

Data is written to the DB and the cache is left untouched until TTL expiry

Question 15

What is the biggest risk of the write-behind (write-back) pattern?

Accepted Answer

Losing writes not yet flushed from cache to the DB if the cache crashes

Answer

Sharply higher latency for each write because it waits for the DB

Answer

Lower hit ratio because new data never enters the cache

Answer

Inability to use TTLs on cache entries

Question 16

What does 'cache invalidation' mean?

Accepted Answer

Removing or marking cache entries invalid when the source data changes

Answer

Periodically clearing the entire cache to free memory

Answer

Rejecting requests that have an invalid cache key

Answer

Encrypting cache data so attackers cannot read it

Question 17

What does a 'purge' operation do in a CDN?

Accepted Answer

Removes specific content from all edge caches before its TTL ends

Answer

Preloads new content into all edge servers ahead of time

Answer

Redirects traffic from a broken PoP to a healthy one

Answer

Deletes old access logs from the edge servers

Question 18

Why are static files often published with names like `app.3f9a1c.js`?

Accepted Answer

So they can be cached with very long TTLs while any content change produces a new URL

Answer

So the browser can download the file without needing HTTPS

Answer

So the CDN can compress the file before sending it to users

Answer

So users are blocked from accessing the origin directly

Question 19

Which HTTP header is the origin's main tool for controlling cache behavior in browsers and CDNs?

Accepted Answer

`Cache-Control`

Answer

`Content-Type`

Answer

`Accept-Encoding`

Answer

`Authorization`

Question 20

What does the `Cache-Control: private` directive mean?

Accepted Answer

The response may be stored only in the user's browser, not in shared caches like a CDN

Answer

The response must not be stored in any cache at all

Answer

The response is cacheable only over HTTPS

Answer

The response is stored only in the CDN and not by the browser

Question 21

What is the difference between `no-cache` and `no-store` in `Cache-Control`?

Accepted Answer

`no-cache` allows storing but requires revalidation before use; `no-store` forbids storing at all

Answer

`no-cache` forbids storing entirely; `no-store` only forbids storing on disk

Answer

Both are identical and exist only for backward compatibility

Answer

`no-cache` applies only to CDNs and `no-store` applies only to browsers

Question 22

What role does `ETag` play in revalidating a cached response?

Accepted Answer

It is a version identifier; if it matches `If-None-Match`, the server returns `304` with no body

Answer

It specifies how many seconds the response stays valid

Answer

It specifies which CDN PoP should store the response

Answer

It encrypts the response so caches cannot modify it

Question 23

What is the `s-maxage` directive in `Cache-Control` used for?

Accepted Answer

Setting a separate TTL for shared caches like CDNs that overrides `max-age` for them

Answer

Setting a TTL only for the browser cache that CDNs ignore

Answer

Setting the maximum cacheable response size in kilobytes

Answer

Setting how long the TCP connection between CDN and origin is kept

Question 24

In a pull CDN, how does content reach the edge?

Accepted Answer

The edge fetches content from origin on the first request (miss) and caches it

Answer

The team manually uploads content to edges before any requests

Answer

The origin immediately pushes every change to all edges

Answer

Users share content peer-to-peer among themselves

Question 25

Which scenario is a push CDN better suited for?

Accepted Answer

Large, rarely changing content like video or installer files known in advance

Answer

A site with millions of pages that are each rarely visited

Answer

API responses that change every few seconds

Answer

Personalized pages for each logged-in user

Question 26

Where does a reverse proxy cache like Varnish or Nginx sit?

Accepted Answer

In front of application servers in the datacenter, between clients and the backend

Answer

Inside the user's browser as part of a service worker

Answer

Behind the database, between the storage engine and disk

Answer

Inside the application process as an in-memory HashMap

Question 27

What is the main problem with a local (in-process) cache in an app running on many servers?

Accepted Answer

Each server has its own copy, so caches diverge and the overall hit ratio drops

Answer

A local cache is always slower than a distributed one because it crosses the network

Answer

A local cache cannot have any kind of TTL on entries

Answer

A local cache can only be used for string data

Question 28

Which of these is an example of a distributed cache?

Accepted Answer

A Redis or Memcached cluster shared by many app servers

Answer

A global variable in a Node.js server process's memory

Answer

The `localStorage` store in the user's browser

Answer

The internal InnoDB buffer pool in MySQL

Question 29

When does 'eviction' happen in a cache?

Accepted Answer

When cache memory is full and entries must be removed to make room

Answer

When a user sends a `DELETE` request to the API

Answer

When new data is inserted into the database

Answer

When the network connection between the app and cache drops

Question 30

Which entry does the LRU (Least Recently Used) policy evict?

Accepted Answer

The entry that has gone the longest since it was last accessed

Answer

The entry that has been read the fewest times overall

Answer

The entry inserted earliest, regardless of access

Answer

The entry that occupies the most memory

Question 31

When does LFU (Least Frequently Used) outperform LRU?

Accepted Answer

When a few keys are consistently popular and a one-off scan should not push them out

Answer

When access patterns are fully random and uniform across all keys

Answer

When all entries have the same very short TTL

Answer

When the cache holds only one key at a time

Question 32

What is a CDN cache key typically built from by default?

Accepted Answer

The request host, path, and query string

Answer

The user's IP address and browser type

Answer

The body content of the origin's response

Answer

The exact request timestamp in seconds

Question 33

What does the `Vary: Accept-Encoding` header tell a cache?

Accepted Answer

Keep separate response variants based on the request's `Accept-Encoding` value

Answer

Cache the response only for clients that support gzip

Answer

Always compress the response with gzip before storing it

Answer

Strip the `Accept-Encoding` header from the request

Question 34

What is 'negative caching'?

Accepted Answer

Caching 'not found' or error results briefly so repeated requests do not hit the origin

Answer

Removing cache entries that were requested too many times

Answer

Storing data with a negative TTL so it never expires

Answer

Rejecting all requests when the cache is unavailable

Question 35

What is a 'cold cache'?

Accepted Answer

A cache that is empty or sparse, e.g. after a restart, so most requests miss

Answer

A cache that stores its data on disk instead of in memory

Answer

A cache that only holds rarely used data

Answer

A cache located in a different region from its users

Question 36

What does 'cache warming' do?

Accepted Answer

Preloads frequently used data into the cache before real traffic arrives

Answer

Tunes the hardware temperature of cache servers for better performance

Answer

Automatically increases the TTL of all keys

Answer

Moves rarely used keys to other nodes

Question 37

What is the 'cache stampede' (thundering herd) problem?

Accepted Answer

When a popular key expires and many concurrent requests miss and flood the DB at once

Answer

When the cache grows so large that the server runs out of memory

Answer

When several cache servers write the same key with different values at once

Answer

When many users download a large file from the CDN at the same time

Question 38

What is the benefit of adding random 'jitter' to key TTLs?

Accepted Answer

Preventing many keys from expiring at once and causing a sudden wave of misses

Answer

Improving security by making cache contents unpredictable to attackers

Answer

Reducing memory usage by randomly compressing entries

Answer

Evenly distributing keys across different cache nodes

Question 39

What problem does a 'hot key' cause in a distributed cache?

Accepted Answer

A large share of traffic concentrates on one node and overloads it

Answer

The key expires quickly because each access reduces its TTL

Answer

The key is automatically replicated to every node in the cluster

Answer

The key's value becomes too large to fit in memory

Question 40

Why is data sharded across nodes in a distributed cache?

Accepted Answer

To grow memory capacity and throughput beyond a single machine

Answer

So every key is copied to all nodes and no data is ever lost

Answer

To remove the need to set TTLs on keys

Answer

So the cache can directly replace the primary database

Question 41

What is the problem with distributing keys via `hash(key) % N` when the node count N changes?

Accepted Answer

Almost all keys map to a different node, effectively emptying the cache

Answer

Only keys of the removed node move and the rest stay in place

Answer

The hash function must be redesigned for all keys

Answer

New nodes cannot receive data until the whole cluster restarts

Question 42

How does consistent hashing work?

Accepted Answer

Nodes and keys are placed on a hash ring, and each key goes to the next node clockwise

Answer

Each key goes to whichever node currently has the lowest CPU usage

Answer

All keys are split among nodes in alphabetical order

Answer

Each key is sent to a random node and recorded in a central table

Question 43

What is the purpose of replicating cache nodes (e.g. a primary with replicas)?

Accepted Answer

Keeping the cache available if a node fails and spreading read load

Answer

Multiplying total cache memory capacity by the number of replicas

Answer

Removing the need for a primary database for durable storage

Answer

Guaranteeing that no data in the cache ever becomes stale

Question 44

What does 'edge computing' mean in CDNs like Cloudflare Workers or Lambda@Edge?

Accepted Answer

Running custom code on edge servers near the user, before or after the cache

Answer

Moving the entire primary database into users' browsers

Answer

Processing all requests in a single central datacenter only

Answer

Compressing files at the origin before sending them to the CDN

Question 45

Which of the following is NOT a benefit of using a CDN?

Accepted Answer

Guaranteeing instant consistency of dynamic data across all users

Answer

Lower latency for users far from the origin

Answer

Reduced bandwidth and processing load on the origin

Answer

Absorbing part of DDoS attacks at the network edge

Question 46

What is the biggest advantage of the browser cache over other layers?

Accepted Answer

On a hit no network request is sent at all, so it is the fastest possible response

Answer

The server can purge its contents whenever it wants

Answer

It is shared among all users of a site

Answer

It automatically encrypts sensitive data

Question 47

What does a database buffer pool (e.g. in PostgreSQL or InnoDB) cache?

Accepted Answer

Data and index pages read from disk, kept in memory

Answer

Full HTTP responses that the app returned to users

Answer

Logged-in user sessions and their tokens

Answer

Static files such as images and scripts

Question 48

In cache sizing, what does 'working set' mean?

Accepted Answer

The set of data actively accessed within a given time window

Answer

All of the data stored in the primary database

Answer

The number of app servers connected to the cache

Answer

The data that is currently being written

Question 49

Why should a system be designed to work (albeit slower) without its cache?

Accepted Answer

Because a cache can be emptied, evict data, or go down, and it is not the source of truth

Answer

Because caches always return incorrect data

Answer

Because using a cache in production is not recommended

Answer

Because the cache is only active in development environments

Question 50

What do CDNs often use Anycast for?

Accepted Answer

Announcing the same IP from many PoPs so network routing sends users to the nearest one

Answer

Sending one response to all users simultaneously like a broadcast

Answer

Encrypting traffic between edge and origin

Answer

Synchronizing cache contents across all PoPs