I have an extension to this question. I have the need for seperation of 
data not so much because of multi-tenancy and security but because the data 
is logically separable and would have a huge impact on performance if I can 
control the volume of data the queries have to refer to. Currently I have 
three solutions all with problems. 

I can use separate databases as that's the best way to visualise the 
logical "partitions" I'm talking about but I have thousands of "projects" 
that need to be partitioned which would create scalability issues within 
the same server. 

Then the second is to have separate sets of classes for each partition 
 with some naming convention for tagging the classes against each 
partition. Both of these solutions will add overhead in terms of managing 
schema and I'm keen to find out how many classes OrientDB can support in a 
single DB. 

The third option of using a cluster within each class for each project 
seems the most natural but fails since we cannot have cluster level indexes 
in OrientDB. Will the performance benefits of using clusters be lost by not 
having cluster level indexes? Any advice will be highly appreciated. This 
is a very common scenario in cloud based and/or high-scalable systems 
design. Ideally I need a light weight partitioning/clustering mechanism 
that does not require you to make compromises.

On Tuesday, October 20, 2015 at 8:44:31 PM UTC+5:30, scott molinari wrote:
>
> Tenant data separation is a standard and huge main security concern, when 
> dealing with cloud computing infrastructure and that same concerns has to 
> be covered by the database as well. It isn't a specific need, it is a 
> necessity.
>
> Scott
>

-- 

--- 
You received this message because you are subscribed to the Google Groups 
"OrientDB" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to