Multiagent systems, being distributed and autonomous, provide the potential for more efficient, component-based development approaches. The massive parallelism intrinsic to these systems promises greatly enhanced performance. However, the price of these potential benefits is often paid in survivability. Distributed multiagent systems have typically failed to provide reliable, assured performance even in well-controlled environments. DARPA challenged the team to achieve this survivability in a way that would convince other potential customers that they could also make their DMAS applications survivable. In response, we developed and applied a generic methodology for survivable DMASs.