Artificial Intelligence for Information-Centric Networks.

The current Internet is an end-to-end communication architecture based on packet switching between terminals. However, with the development of computer networks and increasing proliferation of communication services, the ability of this architecture-based Internet to transmit and process data has encountered challenges, and it has been unable to adapt to the rapid development of today's application requirements. Thus, information-centric networking (ICN) is emerging as a feasible solution and has become a new hot topic for future research in network architecture. The ICN strips content from terminal locations and provides services such as storage and multi-party communication through the publish/ subscribe paradigm. Although the ICN brings new opportunities for network development, there are plenty of problems existing in the ICN that have not been settled. For example, the conflict between polymerizability and mobility in the naming scheme, scalability problems of the name routing, as well as the inefficiency of transmission control algorithms under the circumstance of rich interest are all worth discussing. However, utilizing the computational and analytical features of artificial intelligence (AI) can help the ICN break through the bottleneck of existing technology development. With regard to the issues above, this article introduces and investigates some related technologies of AI in the ICN, that is, naming mechanism, transmission control, and content distribution, and introduces and surveys problems in ICN research. Furthermore, the next development direction of ICN and issues that need to be studied are discussed in this work. [ABSTRACT FROM AUTHOR]