Disasters are exceptional events that are either man made, such as terrorist attacks, or natural, such as earthquakes, wildfires and floods. Disasters create emergency situations and cause physical and social disorder. In these emergency situations, food, water, shelter, protection and medical help are needed, and the effort needed to provide these basic services to the victims must be coordinated quickly via a reliable communication network. A disaster recovery network is used to provide emergency support to both the disaster victims and the crewmembers that are helping the victims, and to provide a communication infrastructure in the disaster affected area. The disaster relief operation also involves searching for and locating the survivors, and then rescuing them. Currently this process involves manual search in the disaster area, which is also time consuming. A novel network architecture called the Portable Disaster Recovery Network is presented in this paper that enables survivors in a disaster area or a non-disaster related search-and-rescue situation to report their locations to a Command Center. This enables first responders to quickly rescue the survivors from these areas. This paper analyzes the performance of random walk models of the movement of survivors in the Portable Disaster Recovery Network.