Check_mk adopts a new a approach for collecting data from operating systems and network components. It obsoletes NRPE, check_by_ssh, NSClient and check_snmp. It has many benefits, the most important of which are:
Significant reduction of CPU usage on the Nagios host.
Automatic inventory of items to be checked on hosts.
The larger your Nagios installation is, the more important get these points. In fact check_mk enables you to implement a monitoring environment exceeding 20.000 checks/min on the first hand.