For a volatile cache, is it better to use memcache or couchbase?
Hi all, I was hoping that you can give me a hand with a situation I'm facing at work. Turns out that I have an application which is currently using memcached for volatile storage of sessions (the session id is the key and then I use a custom Java object as the value). I'm being "adviced" to use Couchbase, but my problem is that I don't know if that's a better solution for the current application. The idea is that at least until now, there is no real storage requirement for that cache (if everything crashes, nevermind) and it's life span is about less than 30 seconds.
One of the things that bugs me is that we have a hard performance requirement to send requests in less than 300 ms (which doubles our current response time) and I don't want to change implementation for introducing a performance impact in my application, just because it's nicer.
What do you guys think? Should Couchbase be a better option, or I should stay with memcache?
Hello,
Couchbase is a great solution for your use case; and it is used a lot for this by customers.
You have probable seen this page on the site that described the difference between a 'simple memcached' versus Couchbase server:
- http://www.couchbase.com/memcached
Also for your application you can choose between the 2 types of 'bucket' :
- Couchbase : behind the cache data are persisted automatically, and you can use the TTL (Time to live) to let the system deal with the eviction
- Memcache : no persistence but better clustering management than standard memcache
300ms... it is not a problem at all, I invite you to look to these benchmarks:
- http://www.slideshare.net/renatko/couchbase-performance-benchmarking
- http://blog.couchbase.com/understanding-performance-benchmark-published-...
Regards
Tug
@tgrall