H-KD: A Novel Query Structure for Multi-dimensional Awareness Information

With the development of Internet, network applications based on awareness information appear increasingly in future Internet, such as context-based routing, content delivery and personalized content recommendation. The full use of awareness information to support those applications makes the fast query of information to be a critical issue. This paper adopts the strategy of "divide-conquer" to propose an effective two-layer data structure named H-KD to search diverse, multi-attribute and massive awareness information. As the first layer structure, Hash Tree organizes the service ID that identifies the type of awareness while the second layer is composed of KD Trees to deal with the specific instances of information. The results of simulation with the real data sets show that H-KD is more efficient and scalable than KD Tree especially when used in the partial query and the range query.