Cache

What is Cache?

Cache is temporary storage that stores the result of frequently accessed data in memory so that subsequent requests are served more quickly
What data should be cached? Mostly used, less frequently changing data is stored in cache. (Eg: Metadata, configuration data)
What should not be cached? Data that is frequently/actively changing.
Cache lies near to requester so that requester need not to go hard-disk to retrieve data. This reduces latency.

Types of Cache

Name	Description
1. Write Thru Cache	Data written to cache and DB both. 1st to cache then to DB. Latency is high Disadvantage: Cache flooding
2. Write Around Cache	Data is directly written on DB. Then later cache updates itself from DB Cache flooding is avoided
3. Write Back Cache	Data is directly written to cache only. Then later DB updates itself from cache Disadvantage: Risk of Data loss(cache may fail)

Where Caches can be placed?

Place

Description

Web Browser Cache

Browser stores local copies of web resources for faster retrieval the next time on Hard-disk.
Ex: Once browser downloads CSS, JS files, images (for a session) there is no need again to download
from same server and can be cached

CACHE-CONTROL HEADER:
HTTP server instructs browser(http client) to cache a particular content for
TTL duration in cache control header. Format: Response Header {cache-control: public, max-age: 45223}

                  @startuml
      
                  box "Client Machine" #LightPink
                  actor User as u
                  participant Browser as b
                  participant "Cache\non Harddisk" as c
                  end box
                  
                  participant WebServer as ws
                  
                  u -> b: HTTP GET google.com
                  b -> c: Has page?
                  c -> b: No
                  b -> ws: HTTP GET google.com
                  ws -> b: Content
                  b -> c: Content, Expiry 
                  @enduml

Between Application Server and Database

Cache Between AppServer & DB

Cache Eviction Policies

No eviction returning an error the memory limit is reached
All keys LRU removing keys by the least recently used first
Volatile LRU removing keys, that have an expiration time set, by the least recently used first.
All keys random removing keys randomly
Volatile random removing keys, that have an expiration time set, randomly
Volatile TTL removing keys, that have an expiration time set, by the shortest time to live first.

Redis vs Memcached

	Redis (Open source implemented in C)	Memcached(Open source implemented in C)
In Memory	y	y
Key-value pair	y (key < 250B, value < 1MB)	y (key < 512MB, value < 512MB)
Speed	less. It does not have any inbuilt datatypes	More. Bcoz it supports datatypes as(string,hash,list,set,sorted set etc)
How internally implemented	Slab Allocator	Encapsulated version of the malloc/free
Scaling	Multi-threaded. Scales vertically. Give more cores, more memory	Single Threaded. Scales horizontally
Cache Eviction policies	Only 1 LRU	6 different policies