A lot of people misunderstand a graphdb with a large-scale graph processing engine, and have asked us before how to run queries etc with Giraph. Nonetheless, giraph can be an OLAP for graphdbs. users having their data stored in their graphdb(s) might want to load their data into Giraph to run analytics.
We can use the API layer offered by Tinkerpop/Blueprints (https://github.com/tinkerpop/blueprints/wiki) on top of many graphdbs. In particular, we can use the REST layer on top of Blueprints offered by Rexter (https://github.com/tinkerpop/rexster).
Quite opportunistically, one option would be to use Faunus's (https://github.com/thinkaurelius/faunus) Rextex-based InputFormat for Mapreduce (https://github.com/thinkaurelius/faunus/tree/master/src/main/java/com/thinkaurelius/faunus/formats/rexster).
The project is a Google Summer of Code 2013 project, and the specifications can be found on our wiki at: https://cwiki.apache.org/confluence/display/GIRAPH/2013#2013-2.Project%3AGiraphintegrationwithTinkerpop