Modeling Node Based Content Repository in graph

[ems eril]

Hi Neo4j ,

   I was hoping I can get some inputs from the community on how to best model a content repository similar to how JCR works but with graphdb. I love to hear what is best way to model these such that

we don't run into scalability issues as the repository grows to billon's of records.

   1) How to model folder and child node relations ship that best scales . We could expect more then million children per folder . It should also be fairly scalable to move things in and out of the folder.

   2) How to best model versioning which includes relationship

   3) How to best model node permission that support also inheritance and overwrites 

   4) How to best model concept of branches where by we can work on parallel copies and be able to converge later

   5) Is there any limitation on node properties if we expect the property to hold large text values such as many pages html content ? And is there any performance lost due to large property sizes or node sizes

  Thanks in advance Emily