Memcache

Store and retrieve data in memory(not persistent) base on specific hash function.

concepts

• Slab: allocate as many pages as the ones available

• Page: a memory area of default 1MB which contains as many chunks

• Chunk: minimum allocated space for a single item

• LRU: least recently used list

ref: Journey to the centre of memcached

we could say that we would run out of memory when all the available pages are allocated to slabs

memcached is designed to evict old/unused items in order to store new ones

every item operation (get, set, update or remove) requires the item in question to be locked

memcached only tries to remove the first 5 items of the LRU — after that it simply gives up and answers with OOM (out of memory)

commands with telnet

• get

• set

  1 2 3 4 5 6 7 8

  set key3 0 0 3                   abc STORED set key3 0 0 4 abcd STORED flush_all

• add: add key or return NOT_STORED if exists

  1 2 3 4 5 6 7 8 9 10 11

  add key1 0 0 3 abc STORED add key1 0 0 3 xyz NOT_STORED add key2 0 0 3 abcd CLIENT_ERROR bad data chunk flush_all

• replace: replace key or return NOT_STORED if exists

• append, prepend

  1 2 3 4 5 6 7 8 9 10 11 12

  set key1 0 0 3 abc STORED append key1 0 0 3 efg STORED get key1 VALUE key1 0 6 abcefg END flush_all

• incr, decr

  1 2 3 4 5 6 7 8 9

  set key1 0 0 3 999 STORED incr key1 1 1000 decr key1 1001 0 flush_all

• delete

• flush_all

• stats

• version

• quit

Run Service

Image used: memcached

Python client: pymemcache

Distributed Caching

Modulo Hashing

• Pros: Balancing the distribution between instances in cluster
• Cons: 1. Loss data if instance down 2. hard to scale

Example

Run 3 instances, expose at port 11211, 11212, 11213

Use python client to set key

Get client instance

pymemcache use Murmur3 hashing

Test with telnet

telnet the third cache server

Consistent Hashing

Scale up / down not affect all the servers on the ring

High Availability

• Repcached: replica data between masters
• KeepAlive: port forword to slave if master down