big fan of memgraph, been utilizing for awhile over neo4j ever since I saw a NASA article detailing their migration effort and the why behind it.
my question is does this multi-tenancy support apply to vector embedding as well? from usage I know it supports vector indexes on nodes/edges, I’m curious whether those embedding properties and vector indexes are isolated and manageable per database in a multi-tenant deployment? any limitations I should know about? context: individual/organization/enterprise knowledge bases being developed need mechanism to promote relevant information up from personal stores into team and from team into corporate.
under Memgraph's implementation of multi-tenancy each database is isolated, meaning that vector search indices are also isolated per database (each db has it's own set of vector indices)
the same limitations apply as to multi-tenancy or vector search (e.g. multitenancy includes some kind of routing, but that's minimal overhead because database is picked per session; importing large vectors can be slower but that's nothing to do with multi-tenancy), there is nothing specific to that combination (to my knowledge)
on your specific use-cases side, I think it's about combining data from multiple databases, that's primarly a multi-tenancy concern, limitation there is that Memgraph can't execute queries across multiple databases in a single query yet (ofc you can manage that on the application layer), there are also options like: cross-database query module -> https://memgraph.com/docs/advanced-algorithms/available-algo..., or Memgraph Zero (totally new product, federated GQL / semantic layer for agents) -> https://memgraph.com/docs/memgraph-zero
big fan of memgraph, been utilizing for awhile over neo4j ever since I saw a NASA article detailing their migration effort and the why behind it.
my question is does this multi-tenancy support apply to vector embedding as well? from usage I know it supports vector indexes on nodes/edges, I’m curious whether those embedding properties and vector indexes are isolated and manageable per database in a multi-tenant deployment? any limitations I should know about? context: individual/organization/enterprise knowledge bases being developed need mechanism to promote relevant information up from personal stores into team and from team into corporate.
Memgraph's CTO here, thanks for the feedback
under Memgraph's implementation of multi-tenancy each database is isolated, meaning that vector search indices are also isolated per database (each db has it's own set of vector indices)
the same limitations apply as to multi-tenancy or vector search (e.g. multitenancy includes some kind of routing, but that's minimal overhead because database is picked per session; importing large vectors can be slower but that's nothing to do with multi-tenancy), there is nothing specific to that combination (to my knowledge)
on your specific use-cases side, I think it's about combining data from multiple databases, that's primarly a multi-tenancy concern, limitation there is that Memgraph can't execute queries across multiple databases in a single query yet (ofc you can manage that on the application layer), there are also options like: cross-database query module -> https://memgraph.com/docs/advanced-algorithms/available-algo..., or Memgraph Zero (totally new product, federated GQL / semantic layer for agents) -> https://memgraph.com/docs/memgraph-zero
getting offtopic xD, I hope the above helps :D