nameddata

Named data networking is an information-centric networking approach in which communication is driven by data names rather than host addresses. It treats data as the primary network primitive and aims to make data retrieval efficient, secure, and resilient to topological changes by enabling in-network caching and name-based routing. The concept evolved from content-centric networking research and is often described in the context of the broader Named Data Networking (NDN) framework.

In Named Data Networking, two types of packets are used: Interest packets and Data packets. An Interest packet encodes the name of the desired content and is forwarded through the network until a node possessing the named Data can satisfy it. A Data packet carries the requested content and is associated with the original name; it is typically signed by the producer to provide data-centric security. Routers maintain three core structures: the Pending Interest Table (PIT) to track outstanding Interests, the Content Store (caches) to hold recently retrieved Data, and the Forwarding Information Base (FIB) to guide forwarding based on name prefixes.

Naming is a central element, with hierarchical or other scalable schemes enabling flexible aggregation, versioning, and segmenting of content. Names are decoupled from host locations, allowing data to be retrieved from any node that holds a match, including intermediate caches, thereby reducing latency and improving bandwidth efficiency.

Security in this model emphasizes data integrity and provenance, with signatures and key management anchored to the data itself rather than to secure channels alone. While promising in concept, practical deployment faces challenges such as name management, privacy concerns related to in-network caching, and the integration of ICN principles with existing IP-based infrastructure. See also content-centric networking and information-centric networking.