Digital twins are beginning to revolutionize many industries in the last decade, offering a plethora of benefits to optimize the performance of industrial systems. They aim to create a continuously synchronized model of the physical system that allows for rapid adaptation to dynamics, primarily to unpredicted and undesired changes. A vast range of industrial domains have already benefited from digital twin technology, such as aerospace, manufacturing, healthcare, city management and shipping. In addition, recent research is beginning to explore the integration of digital twins into computer networks to enable more innovation and intelligent management. One of the building blocks of digital twin technology is the Internet of Things, where wireless sensors and actuators are deployed to provide interaction between the physical and digital worlds. This type of network is complex to manage due to its strong constraints, especially when controlling critical industrial applications, which gave rise to the Industrial Internet of Things (IIoT). We believe that optimizing IIoT will lead to effective integration of digital twins in Industry 4.0. In this paper, we design a Digital Twin Network (DTN) for IIoT where sensors, actuators, and communication infrastructure are replicated in the digital twin to enable real-time intelligent management of these networks. By taking advantage of Eclipse Hono which allows an efficient connectivity for the network devices and Eclipse Ditto for representing the devices states in a digital form and also providing easy access to these states for the DTN. In this way, cognitive network services such as predictive maintenance, sustainability features, network diagnostics, security management, resource allocation, energy optimization can be efficiently integrated and operated in the network lifecycle. We validate the proposed architecture by providing a resource allocation case study where we explain how the Time Slotted Channel Hopping mechanism is exploited in our architecture.