A well monitored data center does not only improve its management, reliability and security, but also helps reduce the energy consumption and the hardware/networking costs. In fact, a constant monitoring allows the early detection of any potential hazards that could lead to a forced downtime or heavy reparations: Either internal risks within the servers and data systems that can manifest through hacker attacks, zombie-servers, network bottlenecks, non-responsive applications and bugs; or external facility risks related to unexpected or sudden changes of environmental factors such as partial power outage, fire or climate abnormalities. Currently, the existing monitoring solutions either focus on only one hazard source (internal or external) or both via external Cloud-based services that may alter the security and privacy of the data as well as may cause network bottlenecks. The main purpose of this thesis is to present a state-of the art concept that allows a centralized monitoring of the servers, hardware and facility of a data center independently of its topology and network infrastructure. The introduced concept takes advantages of both Internet of Things and Edge Computing in order to reduce its load on the data center network while maintaining a high monitoring accuracy. We have implemented a framework overarching the described concept in order to evaluate its effectiveness and influence on the data center bandwidth. We have also simulated a data center server and explored its monitoring possibilities, then compared the results of a traditional server monitoring with those obtained when using our framework. The results were promising enough to encourage the adoption of the presented concept in real data centers. However, future work has to be carried out in order to address the limitations outlined in the concept.