NoSQL combinations - what works best?
wmoss at bu.mp
Tue Aug 9 13:56:20 EDT 2011
Apparently Antirez is currently working on a cluster solution to Redis, but
in the meantime, he's written a good blog post about how to get around the
TL;DR: Redis is light-weight, so just run a bunch of instances on one
machine and ring hash among them. If you need to distribute your load in the
future, just move some of those instances to a new machine.
We've done something like this with good success for things where we need
more ram/cpu than one redis instance can provide. Obviously things like set
intersections don't work across instances, but if you're not using that (or
you make sure your hashing scheme is consistent enough to allow for the
intersections you need), it works great.
On Tue, Aug 9, 2011 at 10:48 AM, Alexander Sicular <siculars at gmail.com>wrote:
> Hmm. Not that I know of. Afaik, the memcache server itself is standalone.
> Any distributed magic happens at the client level via some circular hashing
> voodoo or sharding pixie dust. Keep the client mojo and swap memcached for
> @siculars on twitter
> Sent from my iPhone
> On Aug 9, 2011, at 13:23, Les Mikesell <lesmikesell at gmail.com> wrote:
> On 8/9/2011 12:08 PM, Alexander Sicular wrote:
>>> Outside of legacy concerns I have no idea why someone would select
>>> memcache over redis for a new application today. Redis gives you
>>> everything memcache does plus a boat load of
>>> omg-wtf-where-has-this-been-**all-my-life capabilities.
>> The main point of memcache is that it is distributed and the clients
>> automatically handle failover at any reasonably large scale. How do you
>> arrange that with redis?
>> Les Mikesell
>> lesmikesell at gmail.com
>> riak-users mailing list
>> riak-users at lists.basho.com
> riak-users mailing list
> riak-users at lists.basho.com
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the riak-users