In this article we introduce a new conception of three-dimensional DataSpace, which is physical space enhanced by connectivity to the network. DataSpace is addressed geographically as opposed to the current logical addressing scheme of the Internet. Here, a local area network is replaced by a room, a street, a mountaintop, and so on. Billions of objects populate DataSpace, each aware of its own geographic location. These objects move through DataSpace, and produce and locally store their own data. They can be selectively queried, monitored, and controlled based on their properties. We propose two architectures for DataSpace. We describe mechanisms to use the network as a DataSpace engine in order to perform querying and monitoring operation in a highly scalable way.