The boundary layer exists because of the no-slip condition in real-world flows. Because velocity right at the wall must be zero, there must be some velocity gradient between zero velocity and the free-stream velocity. The part of the flow between the wall and 99.9% of the free-stream velocity is said to be the boundary layer.
If one non-dimensionalizes the Navier-Stokes equations, I believe there is a term that looks something like:
(1/Re)*dv/dy
Because the term in front is so large, the velocity gradient must be large to make it "matter" in the equations. (note: I'm 50% sure of the equation, either way its something like that).
That is the reason that the boundary layer is so small (think <1/2" on your car at 60mph).
Hope that's a start.
edit: Also, you can directly apply boundary layer equations for a flat plate for your simple analysis. For the leading edge of the airfoil, you can use the Hiemenz (sp?) stagnation solution. Then, because the boundary layer is thin compared to the curvature of the airfoil, the flow essentially "sees" that its flowing over a flat plate.
