Projection-free Distributed Online Learning with Strongly Convex Losses