What advantages does click have over how we currently do things?
Well, we're using this product to play around with routing protocol implementations to enhance performance, convergence, stability, and to test new concepts in a real life testbed compared to simulation-only approaches.
For instance, it would be real easy to implement mesh networking for community projects with the WRT using TBRPF, DSR, AODV, or OLSR and to modify these protocols to meet your needs. Implementations are all C++, and modules, which each router is based on, can be written to your heart's content. Once the click router infrastructure is in place, you can change your routing on the fly and test what works best for you. Each of these protocols has its pros and cons.
There is a userlevel driver, and a kernel driver that each can execute a configuration file that stores the actual router. The kernel driver (module)really boosts performance, but needs a kernel patch that is not easily implemented.
Like PolarWolf, I will see how far I get compiling Click for mipsel, against uClibc with the 2.4.20 kernel. If we can pull this off, it will enhance the flexibility and usefulness of the WRT as an experimentation platform by some orders of magnitude.
- Gromit.