Hi, I am currently evaluating OrientDB product as Document database against other solutions. We use Document database because users must be able to add/remove any fields they want into their records. I have recently discovered this DB and seems to be "on paper" a great solution and could fit perfectly to our needs mostly because we can mix relational and document oriented data. I am also interested in hooks we can insert at server side for CRUD operations and record level security. One of our test to evaluate product is related to search performances when we have nearly 1M (1 000 000) documents into the database. I clearly understand that we need an index to have quick search on specific fields but we also need to have average performances for "free" fields added by users (we cannot know them).
When running the search on these 1M documents (one single class and 1M records with nearly 20 fields) we can notice bad performances and strong CPU/memory consumption. It takes nearly 25 sec to have result. 1) How OrientDB works exactly when searching ? Is it mostly in-memory read or filesystem read ? For example, do OrientDB need to load all records in memory for search ? 2) Is there any tuning i could use to speed up searches when i do not have indexes ? In order to have a comparison point, we have tested the same with Mongo and result is returned in nearly 5 sec without noticeable memory/cpu consumption. I understand OrientDB propose a lot more features than Mongo but can we expect to have similar performances and what could the "cost" in CPU/Memory (memory size, numbers of CPU) ? Is it possible ? Regards -- --- 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.
