So one of the things that I've thought about with when I've considered using HBase for RDF storage was how to implement blank nodes. I've always talked about requiring a global lock on the system in order to ensure that if you refer to a blank node on one node in the cluster that it is the same node in another. There are ways around it (using some form of context around the blank nodes preventing the data relating to a blank node from splitting in the first place) but it still seems like an outstanding problem. I'd be interested in this part of your solution.
2009/3/24 Philip M. White <p...@qnan.org>: > On Mon, Mar 23, 2009 at 05:33:46PM -0700, stack wrote: >> Anywhere we can go to learn more about the effort? What can we do in HBase >> to make the project more likely to succeed? > > Right now we don't have anything of value to show you, but we plan to > move on this pretty quickly. We're copying the functionality of using > HBase as the persistent store from another (proprietary) project. > > If you (or anyone else) would like to participate in this development, > let me know. We can work together on this. > > -- > Philip >
