Adaptive web caching views the caching problem as a way of optimizing global data dissemination. A key problem adaptive caching targets is the hotspot phenomenon, where various, short-lived Internet content can, overnight, become massively popular and in high demand.
Adaptive caching consists of multiple, distributed caches which dynamically join and leave cache groups (referred to as cache meshes) based on content demand. Adaptivity and the self-organizing property of meshes is a response to those scenarios in which demand for objects gradually evolves and those in which demand spikes or is otherwise unpredictably high or low.
One of the key assumptions of the adaptive caching approach is that deployment of cache clusters across administrative boundaries is not an issue. If the virtual topologies are to be the most flexible and have the highest chance of optimizing content access, then administrative boundaries must be relaxed so that groups form naturally at proper points in the network.