Today's modern mobile devices (such as smartphones and tablets) present great potential for growth of many novel, powerful, but also highly demanding applications. However, most mobile devices/users operate in isolation from one another, i.e., they are not aware of the presence of other devices in their proximity. There are numerous situations where proximity-awareness (i.e., a device is aware of other mobile devices in its neighborhood) could be used to support spontaneous sharing of resources and information, thereby enabling a variety of new application scenarios. This dissertation presents a software architecture called SPontaneous Information and Resource sharing InfrasTructure (SPIRIT), which allows mobile devices to create, discover, join, leave, and control the sharing of resources and information in an efficient and automatic fashion. Built on top of a group communications substrate called Courier, SPIRIT allows users or devices to express various heterogeneous service sharing paradigms using a cleanly defined Application Programming Interface and a powerful Subscription Language. These service sharing specifications are used by a signaling protocol in SPIRIT that allows the service-seeking clients to tell the service provider and intermediate nodes about the specific Quality of Service (QoS) required in the requested service stream. SPIRIT provides additional QoS mechanisms like mobility-aware admission control, resource reservation, management and policing, in order to establish and maintain QoS along the multi-hop paths between the service provider and the client in mobile environments. While the shared services can include various types of resources (e.g., network connections, CPU) and information (e.g., database entries), this dissertation focuses on sensor information, i.e., the information collected by the various sensors found in modern mobile devices (e.g., GPS, acceleration, image, light, pressure, sound, etc.).