Cognitive Computing Resource Management for a Ubiquitous Wireless Access

摘要

In the near future many radio access technologies (RATs) will coexist to offer a vast variety of user services at different quality levels. Software-defined radio (SDR) facilitates to flexibly choose the RAT by means of a software-reconfiguration of the SDR terminal. Cognitive radio automates this process, being an ambient aware, intelligent radio resource management system. We extend its scope in that we introduce the cognitive computing resource management. Our cognitive radio system scans the computing environment, decides whether to reconfigure a candidate SDR mobile terminal or not, and learns from the effects of these decisions. The simulations, which model a simplified though realistic case study, show that two simple computing resource management algorithms outperform an algorithm that is unaware of the present computing environment. The paper concludes that a cognitive computing resource management is necessary for providing a ubiquitous wireless access where the user demands are satisfied as a function of the currently available computing resources of their terminals.