From our first analysis, applying the thrust through the container seems doable; the containers are secured to each-other and the ship via mechanical "twistlocks", which are designed to handle not just ship loads, but also trucking loads (such as a 2g hard braking from a semi). The roll loads are actually the more difficult design challenge.

Regarding kites, we looked at those quite a bit. The challenge with those is that kites tend to be best for when the wind is coming from behind you or crosswind. For a container ship traveling at high speed, the kites would act more as a parachute and slow you down (even if you were extracting energy from them).