Deep-space missions are heavily constrained by the amount of payload mass the launch vehicle can carry. Furthermore, the amount of payload mass the launch vehicle can carry is limited by the delta-V losses of escaping both Earth's gravity well and its atmosphere. Instead of launching the propellant mass to be used for trajectories to deepspace, if the propellant can be delivered in-space, the vehicle may carry a significantly larger payload from the surface of the Earth to the destination. Such an architecture is a paradigm shift for space exploration, enabling spacecraft to fly to the furthest reaches of the Solar System with more mass and/or in less time. An international team of sixteen students met at the 2017 Caltech Space Challenge to design Lunarport: a station which provides vehicles traveling to destinations around the solar system with propellant created from water ice extracted at the lunar south pole. A complete system concept design and architecture was produced, entitled 'Ice Rush', which leverages mostly TRL 6+ technology and is capable of refueling crewed Mars missions by 2032 at a total cost of $17B. A detailed analysis of in-situ resource utilization methods, propellant depot design, lunar site selection, and prospects for decreasing costs/increasing payloads of future deep-space missions is included. With the Ice Rush architecture, launching an SLS Exploration Upper Stage (EUS) to Lunarport's L1 depot and refueling, the payload mass may be tripled for a mission to Europa or doubled for a free-return trajectory to Mars. A solar electric space tug concept is also presented, which would triple the Mars freetrajectory payload mass using Lunarport.