I don't consider the variation within dogs to be micro-evolution. I don't think the extremes within dogkind are part of some trend leading to something else millions of years down the road. They're simply the outer parameters (in this case in terms of size) of variation within a particular kind. I would propose that variation within kind is not micro-evolution, but a form of conservation. Allowing for some variation within kind promotes survival of the kind. For example, long haired wolves might survive an ice age that would kill short haired wolves. Smaller species members might survive a food shortage that would starve the larger ones.
It should also be noted that man has pushed against species boundaries to produce extremes such as Chihuahuas and Great Danes, but we only got that far because dogkind (for lack of a better term) had the genetic capacity for those size extremes already. We can't push past them. We can't breed dogs with wings, or horns, or other attributes not possible within the dog genetic code.
To assume that over time dogs could "evolve" past their species boundary is just that, an assumption with no basis in the observable world of science.
I suppose that I'm only assuming that Earth history will go on, so I guess I am only making an assumption that dogkind would evolve into something seriously different given enough time.
Since Creationists have no described mechanism to their "species boundary" theory, then I will further assume that it is (to quote Dataman) "just a guess".
So what you are saying is that natural selection is happening. The fact that different variants of a species are not equally capable of surviving and reproducing is the very essence of natural selection. Also, to assume that over time dogs could NOT evolve past their species boundary is also an assumption with no basis in the observable world. It depends on what you assume the species boundary to be. For example, is it within the species boundary of certain bacteria to be able to survive in the presence of chemical compounds that kill most bacteria? I think, considering that there's no clear and distinct way to differentiate species of organisms that reproduce asexually, you'd have a hard time making an objective determination of the species boundaries of bacteria in the first place.