Читать книгу The Smart Cyber Ecosystem for Sustainable Development - Группа авторов - Страница 48

Current networks are characterized by their distributed nature, as each node (router/ switch) has the ability to view and act on the system partially and locally. Thus, global learning from network nodes that have a holistic view on the system will be very complicated. Further, current network designs impose significant limitations on network performance, especially under high traffic conditions. Consequently, the increasing demand for reliable, fast, scalable, and secure networks can adversely affect the performance of existing network devices due to the need to deliver a large volume of data both in the network infrastructure and devices. Current network devices lack the flexibility to handle different types of packets that may carry different contents due to the basic implementation of hard-wired routing rules. In addition, the networks that form the backbone of the Internet must be able to adapt to the changing conditions without needing much effort for hardware and software adjustments.

In order to reach a solution to the above discussed limitation issues, the rules for data processing must be implemented through software modules and not embedded in the hardware. This approach enables network administrators to have more control over the network traffic, and thus can greatly improve the network performance and effectively use the network resources. This innovative approach is called Software-Defined Network (SDN) [9].

SDN was released as open source software in 2008 with the OpenFlow project at Stanford University. It decouples the control and data planes in routers and switches, allowing the underlying infrastructure to be separated from application and network services. Thus, SDN separates the decision-making modules about where traffic is sent [the control plane (CP)] from the underlying systems responsible for forwarding the actual traffic (the data plane). Network resources are managed by a centralized controller which performs as the network operating system. The controller can dynamically program the network in real time. It collects information about network status and operation details. Therefore, the controller can globally detect available network resources and requirements. This paradigm creates a global view of the entire network, enabling global automatic management and control without needing to configure devices individually. The SND technology has several advantages:

 Efficient utilization of network resources.

 Enables development of programming-based solutions for network configuration and management.

 Provides a perfect ecosystem for ML paradigm and intelligent applications.

 Simple and improved network management, control, and data manipulation, since network administrators can remotely alter the network configuration and operation as response to dynamical changes in the network.

 High speed, through optimal handling of the traffic load.

 Adopts the virtualization technology, which allows running multiple applications over the same shared hardware.

Combining AI and SDN has been attracting researchers in recent years to develop network management and operation mechanisms. The SND architecture provides centralized control of network policies and enables administrators to effectively overcome problems with ML methods.