The Moving Cloud: Predictive Placement in the Wild

As available bandwidth improves, the role of latency as the system bottleneck becomes even more pronounced. This is particularly concerning as cellular networks become the primary avenue of connectivity for many users in developing regions. This paper proposes the moving cloud, an architecture for the proactive delivery of data. The moving cloud leverages route fingerprints in individual mobility with users' contextualized behavior of data access. The moving cloud trades bandwidth and storage for latency by predictively placing content where it needs to be in the near future. This is enabled by a novel mechanism for augmenting location prediction algorithms with time bounds, allowing for a timely delivery of data. This paradigm enables a number of networking scenarios ranging from bulk data access and mobile resource augmentation to personal content distribution.