THE CONCEPTUAL MODEL OF INTELLIGENT MANAGEMENT OF NETWORK RESOURCES IN THE IMPLEMENTATION OF THE SDN/NFV PARADIGM

Abstract

B a c k g r o u n d . With the development and spread of network technologies, the problem of effective management of networks is becoming increasingly relevant. One approach to solving this problem is to use network virtualization, such as software-defined networking (SDN) and network functions virtualization (NFV), which allows to use of non-specialized hardware, including hardware on the x86 architecture, and makes networks more efficient due to the optimal use and distribution of resources. Due to this, it becomes possible to use own models and methods for managing networks. The purpose of the work is to find and develop a conceptual model of intelligent management of network resources in software-defined networks, which, unlike the existing ones, can dynamically allocate resources depending on the needs of the network and can adapt to possible minor discrepancies between the received data and real needs under the conditions of external and internal destabilizing factors. M e t h o d s . In this paper methods of neural networks, fuzzy logic methods as well as methods of computer simulation are used. R e s u l t s . In this study, it is proposed the conceptual model of network resource management based on a hybrid neural network using fuzzy logic and fuzzy output inference (ANFIS), which allows for dynamic response to network needs and is less sensitive to the rapid obsolescence of received data regarding its needs. The model was tested on a simulated communication network in the MATLAB environment in two scenarios with low and high network load. It showed a good result that allowed the virtual devices to handle requests without overloads and have a certain margin of computing resources for efficient work and possible minor growth of network needs. C o n c l u s i o n s . The developed model of network resource management in the MATLAB environment showed high efficiency, which proves the necessity and relevance of further research on using fuzzy logic in network management.