How to partition large amount of connection data ?

[Ashish Jain]

I am confused about how to partition the data for network connections. If its a simple type ahead, documents can simply be partitioned to different machines using document id ranges. But how can I partition the connections data ? Say the data is large and needs to be partitioned horizontally to two machines:

Machine 1: Document 0 to 99
Machine 2: Document 100 to 200

Say documents are user ids in the network

Now, what if document 75 has a connection to document 150 ? When uid=75 is searching for something, how will cleo verify document 150 qualifies for the query or not ?

[Ashish Jain]

Hi Jingwei, can you please give some hints for this problem