The benchmark highlights the value prop to the end user who can now use a "local program" ie an embedded database and get the perf wins over Redis - without worrying about data management (backups/replication/what-not).
Read path don't invoke any RPCs (even on startup or a cache miss).
Writes need RPCs - since they have to be propagated to the (many) readers.
Benchmarking a local program vs an AWS hosted service to claim you’re faster than Redis?
Reads are served locally but all writes go to some hosted cloud service?