Memcached Cluster – better to have 3 smaller instances or 2 mediums ones

You really should use a cache server in front of your web servers. I recommend Varnish-cache. We use it at work with the largest and busiest website in scandinavia. We replaced 13 highly loaded Squid boxes with 1 Varnish box, and 1 for spare.

I benchmarked a simple app on my private website, and it went from 9 requests a sec to over 2000.

You decide how long it keeps things in memory, you can do a till the end of time and then just send a http purge request to the cache server when the data changes.

Andrey,

To answer your question I would refer you to the snippet below. Performance wise memcache client library would still be doing two internal lookups regardless of whether you have 1 or more instance.

But having only one instance might hurt because it would take sometime for the backup to warmup when it comes online.

http://www.linuxjournal.com/article/7451?page=0,1 A request to get/set a key with a value requires that the key be run through a hash function. A hash function is a one-way function mapping a key (be it numeric or string) to some number that is going to be the bucket number. Once the bucket number has been calculated, the list of nodes for that bucket is searched, looking for the node with the given key. If it's not found, a new one can be added to the list.

So how does this relate to Memcached? Memcached presents to the user a dictionary interface (key -> value), but it's implemented internally as a two-layer hash. The first layer is implemented in the client library; it decides which Memcached server to send the request to by hashing the key onto a list of virtual buckets, each one representing a Memcached server. Once there, the selected Memcached server uses a typical hash table.