Heartbeat mechanism is widely used in the high availability field of monitor network service and server nodes. In this paper, we design a distributed heartbeat mechanism, which consists of one Master node and multiple Standby nodes, all running the same process of heartbeat but in different states. We experimentally demonstrate that the mechanism improves the scalability over that of the dual computer nodes and supports nice fail-over. Moreover, the process of the heartbeat only needs to maintain few states, which dramatically reduces the overheads of the system and simplifies the management of the nodes. When there are only two nodes in the network, the mechanism will degenerate to dual computer heartbeat. In current version, the mechanism only supports the switch-based architecture; it is easy to make it supporting other architectures such as Ring, etc.