I'm not aware of open-source projects making use of Voldemort, though maybe the community has better insight on that.
The use cases for Voldemort are the following:
- Voldemort Read-Write:
- Source of truth data (typically tuned with a quorum style consistency).
- Derived data (potentially tuned with looser guarantees).
- All in all, you can use Voldemort RW for most of the use cases that vanilla Cassandra (or vanilla HBase to a lesser extent) are used for. That being said, the ecosystem around these other projects is stronger, so there aren't Voldemort equivalents to Cassandra's CQL or HBase's Phoenix, for example.
- Voldemort Read-Only:
- Bulk loading data computed on Hadoop.
- Ultra fast online serving of derived data.
- This use case is probably where Voldemort shines the most, since it is very mature by now, and also because there aren't many solid contenders (AFAIK) for this in the open-source space.
Another advantage of Voldemort (both RW and RO) is that it has built-in (and pretty solid) support for multi-data center deployments.
Hope that helps. Feel free to ask if you have more questions.