mongodb as cache leyer

[mongoUser1]

Hi ,

we are using MongoDB currently as replica Set to Sync data between our data centers .
to speed up performance we need to implement some kind of cache layer .
we currently considering memcahed & mongodb .
the data stored in the cache is key-value type , with small size strings .
the cache should be distributed at the data center between computers there .
we have mongo version 2.0 and not 2.2 where the TTL collections exists , so we implemented it by a thread that remove old entries after the TTL .
we expect the cache size to be ~10 milion records .
and we expect the cache to be queried ~1000 times a second .
will the mongoDB holds in this kind of hit rate , where the same collection will be scaned very often with reads , and written too (in smaller scale) , and deleted from ?
won't the memcached solution per data center be better in performance?
this way the primary won't be written at all (just from our MS-SQL DB , not from the side of the clients),
and the queries to my understanding will be faster (am i right?)
i hesitate to add another layer of the memcached , and if the mongoDB will handle it i gladly will choose it .

thanks.

[Max Schireson]

~1000 per second over 10 million records sounds well within range of course hard to say for sure without more details, but plenty of folks do more than that even on a single box.

2.2 handles mixed read/write workloads better; i'd consider that upgrade unless the writes are extremely infrequent.

-- Max

--
You received this message because you are subscribed to the Google
Groups "mongodb-user" group.
To post to this group, send email to mongod...@googlegroups.com
To unsubscribe from this group, send email to
mongodb-user...@googlegroups.com
See also the IRC channel -- freenode.net#mongodb