This will not buy you much. If you have MySQL on the backend, it is
useless idea to replace storage engine by Memcached.
MySQL itself has a bunch of optimization options that effectively turn
disk reads into memory reads . In your case I would guess that following
tuning would help:
1. Increase MySQL query cache size
2. increase key_buffer_size (normal value is 512K).
MyISAM is also faster than InnoDB, if your application logic permits
non-transactional DB. MEMORY engine is probably not a good idea since
you can loose all your data.
Angelo McComis wrote:
I am running a fairly large TinyDNS infrastructure. That (for those that
don't know) uses a MySQL backend as a datastore. I'm looking for ways to
improve the performance of this thing. MemcacheD looked appealing as it
could theoretically plug in as a storage engine and turn disk reads into
memory reads for frequently resolved DNS queries.
Is there a better solution?
--
Alexander Zaitsev
Engineering Manager
AMG Lab Sàrl
