This looks quite interesting! - http://lkml.iu.edu/hypermail/linux/kernel/1708.0/02224.html Not being just a traditional file system, I wonder if we could leverage some of its specific features like COW and the log structure to avoid such redundant operations from our side. For example Lucene's indexes work hard to re-implement a similar schema on top of a general-purpose filesystem. I'd assume that the filesystem has some oppoprtunities to do it "better". Might be of interest for Our off-heap and CacheStores as well. Thanks, Sanne