Abstract-Exploration of extreme environments, including caves, canyons and cliffs on low-gravity surfaces such as the Moon, Mars and asteroids can provide insight into the geological history of the solar system, origins of water, life and prospect for future habitation and resource exploitation. Current methods of exploration utilize large rovers that are unsuitable for exploring these extreme environments. In this work, we analyze the feasibility of small, low-cost, reconfigurable multirobot systems to climb steep cliffs and canyon walls. Each robot is a 30-cm sphere covered in microspines for gripping onto rugged surfaces and attaches to several robots using a spring-tether. Even if one robot were to slip and fall, the system would be held up with multiple attachment points much like a professional alpine climber. We analyzed and performed detailed simulations of the design configuration space to identify an optimal system design that trades-off climbing performance with risk of falling. Our results show that with increased number of robots, climbs can be performed faster (through parallelism) and with less risk of falling. The results show a pathway towards demonstration of the system on real robots.