Graph representation in Sparksee

[oj]

Hello Sparksee team,

I'm trying to understand a few things regarding the way graphs are represented (in memory and on disk) in Sparksee.

When a query is issued to the database, my understanding is that the graph (nodes/edges/properties) is loaded into memory using a bitmap structure for efficient query processing. 

How is a graph in Sparksee represented on disk? How is this representation optimized so that for a particular query, the number of disk accesses are minimized?

[sparkseegdb]

Hello!

You can find more information about how Sparksee uses bitmaps structures in this paper: http://dl.acm.org/citation.cfm?doid=2351476.2351489

Hope it is helpful!

Sparksee team