- Currently each block is written to RDFS in a sequential fashion. - Making each `write` request parallel would increase overall throughput. - These facts are similarly true for a `delete` request. - To do so, we need to specifically change implementation in the `Write` and `Delete` commands in `rdfs-client`.