This counter counts specifically the read-type I/O requests served by this node's MSCP server for remote cluster members. Comparing this to the write-request count tells the admin the read/write balance of the storage traffic this node is serving on behalf of others. For example, a remote node running mostly reporting jobs that scan large datasets would drive this read counter up much faster than the corresponding write counter.