Turbulence observations over recent decades show that mixing intensifies toward the seafloor, implying that abyssal upwelling cannot occur broadly through the interior as classically assumed, but must instead be confined to thin bottom boundary layers along sloping topography (Ferrari et al. 2016; McDougall & Ferrari 2017). The Boundary Layer Turbulence and Abyssal Recipes (BLT) experiment set out to test this paradigm in a submarine canyon on the steep boundary of the Rockall Trough. Moored turbulence measurements, microstructure surveys, and a near-bottom dye release delivered the first direct evidence of vigorous diapycnal upwelling (~100 m day⁻¹) confined to the bottom boundary layer (Wynne-Cattanach et al. 2024; Naveira Garabato et al. 2025), driven by energetic turbulence phase-locked to the tidal cycle. I will close by presenting a dynamical framework that explains the observed upwelling: using an idealized box-canyon model and realistic simulations, I show that the canyon acts as a resonant cavity for standing internal Kelvin waves, whose bottom-intensified shear supplies the turbulence and upwelling — linking basin-scale overturning theory, canyon-scale wave dynamics, and centimeter-scale mixing.