Reputation: 10716
Given a graph where there's nodes of type users and social identities, and relationships of type knows and has_social_id, like this:
User A has_social_id FB1
has_social_id Another
User B has_social_id FB2
User A knows User B
Foreach user entity only the email, a boolean property and 2 date properties are stored, for the socialid entity only an integer property and a date property are stored, and that each user has on average 100 friends and 1.1 social identity.
Do you think the community edition can handle about 80000 users, each with ~ 100 friends, with reasonable performance without the need for clustering only increasing a single machine's capacity?
Upvotes: 1
Views: 321
Reputation: 1385
Yes, this is a graph of 80,000 people with 800,000 relationships. Very much within the capabilities of a single Neo4j instance (which deals with many billions of nodes/relationships with ease).
I've seen examples of Neo4j running at about 60% of the Facebook social graph, so if you're anywhere between 1% and 60% of Facebook, then Neo4j is proven to work. At 61%, you're breaking new ground :-)
Jim
Upvotes: 6