Clarification of SB Tree algorithm not working on a distributed database


Can someone expand on this statement:
When running a distributed database, the SB Tree index algorithm is not supported.

Found here:

What exactly does this mean and what needs to be done or known to deal with it?


Hi @eric24

SBTree Bonsai structure is an optimized data structure to represent edges (specifically designed to address situations where you have a lot of edges on the same vertex).
Unfortunately, the distributed protocol does not support the replication of this structure, so when you run distributed, OrientDB falls back to embedded ridbags (ie. it stores edge RIDs inside the vertex).
You don’t have to do anything, OrientDB will do the conversion for you




OK, but what does that do for vertexes that have lots of edges (“super nodes”)? Without the SBTree Bonsai structure, what is considered a useful/workable limit to the number of edges per vertex?

Also, is support for the SBTree Bonsai structure in distributed mode on the roadmap, or is this considered impossible/impractical? Or is it not important for some reason?