To the servers that are installed in-house yes. Not in other locations. ClusterFS can operate on different levels. distributed (similar to RAID0) replicated (similar to RAID1) distributed replicated (similar to RAID10) striped (pieces of data are spread across different locations, for load reducing) distributed striped (same as above but also with redundancy)
I am using raidz3 underneath Gluster, so redundancy is taken care of on ZFS level mostly. As big as you budget can be stretched. Mostly curiosity of different approach to data handling and management. I do not think about profits at any stage yet. Maybe in year time when project mature.
Also side note. After upgrading to 1.2.1 "host unreachable" has been fixed. It is "online" now.**