This has been discussed many times here. Maybe you need to have a look at the archive. There are many existing solutions for this problem, of course. They all have their advantages and disadvantages, when seen from an amateur viewpoint. However you need to understand that the sending system has no information about "the route being unreliable" and the receiving system has no way of making the sending system try other routings. That is why a working automatic routing protocol is always much more complex than RIP (which works like the method used by NET/ROM) There is little point in rehashing time after time how much better everything could be, when there is not going to be a change anyway. Rob