RAM isn't a cure all for big data applications and presents its own set of challenges. The largest issue is with traversal and data access latencies simply beca

RAM isn't a cure all for big data applications and presents its own set of challenges. The largest issue is with traversal and data access latencies simply because of the virtual memory management overhead that is required. I've run into this problem with our btree search indexes on bare metal (~96GB), some more reading: https://github.com/johnj/llds https://github.com/johnj/llds