The OrderOne Networks Mesh Network Routing Protocol was designed from the ground up for easy integration. It has no OS calls of any sort built in. Instead it hooks to your OS or application via approximately 100 lines of 'C' code provided by OrderOne Networks.
Since the protocol was written entirely in 'C' it may be compiled for virtually any 32bit CPU and OS combination. It uses no floating point math and is completely endian independent. The compiled size of the protocol library depends somewhat on the CPU/OS combination but is typically around 200K.
Since the runtime memory and CPU requirements are adjustable, the OrderOne Networks Mesh Network Routing Protocol may run on anything from a low power embedded device to a desktop or server.
The following pseudo code illustrates how to integrate the protocol into an OS or simulation. The OrderOne Networks Mesh Network Routing Protocol also supports an event driven model without the sleep or the loop. Additionally, there is an internal routing table that may be used instead of updating an external routing table.
The function parameters are mainly omitted, but the order and the function names are correct. OON_* are OrderOne Networks Mesh Network Routing Protocol functions. OS_* are functions provided by the Integrator, a simulation or the OS.
And yes, this really is all you need to start building large, mobile mesh networks.