Despite the rapid growth in popularity and hardware capacity in mobile devices, they suffer from resource poverty, which limits their ability to meet increasing mobile users' demands. Computation offloading may give a prominent solution. But it relies on the connection to the remote cloud and may fail in situations where there is poor or no connectivity. Cloudlet was introduced to cover this problem, but mobile users miss free mobility when using cloudlets. Offloading to the cloud or cloudlet is not always the preferred solution. An alternative is to utilize the nearby mobile devices as local resource suppliers and pull their capabilities as a mobile device cloud. In this paper, the authors present such an approach known as ad hoc computing as a service (AhCaaS) model for computation offloading in an ad hoc manner by connecting to nearby mobile devices. They define a multi-attribute selection strategy to determine the optimal computation offloadee. They evaluated the proposed model, and the result shows that AhCaaS reduces execution time, battery consumption, and avoids task reassignment.