*3.5. Complexity*

It refers to the difficulty to resolve computational problems analyzing the algorithms implemented in the WDs and the use of resources as memory, storage, energy, time, etc., [39]. The challenge in complexity is to implement efficient algorithms for wireless systems that permit to reduce the power consumption and the size of devices while decreasing its manufacturing cost. However, the simplicity must ensure the *quality of service (QoS)* and privacy.

## **4. Performance Metrics Study**

Metrics of the network layer are fundamental because they show the performance and usefulness of a routing protocol. Each routing protocol is designed for specific applications and certain scenarios. These metrics indicate how the bandwidth use is affected by the overhead of the routing protocol in use. In addition, the availability of effective routes and the ability of the network for self-configuration show the capacity of the protocol to recover from topology changes [40]. Recovery times have an impact on the latency in the network, and even though the networks conform with different technologies, it is highly essential to understand and evaluate the performance metrics as shown in [41]. This analysis allows network customization to improve its different aspects and provide better communication.

Therefore, using these metrics the proactive and reactive protocols can be compared. To perform such comparison a simulated and a real scenario were implemented. The next section describes the simulation.

Figure 2 describes performance metrics in a general way. The measurement and control of these performance parameters provide the network optimization because some parameters influence others, allowing the information in a wireless network to be delivered reliably.

**Figure 2.** General performance metrics.

Sensor networks usually have similarities, so it is necessary to know some comparison criteria that allow us to decide which of them is the most suitable to implement and thus, solve our problems. The most important and useful metrics for comparing sensor networks are the following:


• *Packet loss:* It occurs when information packets do not arrive in the expected way. Some of the causes can be damaged hardware. Hardware capacity (bottlenecks): when some component of the network cannot assume a certain amount of traffic, thus encouraging data flow. Congestion in the network: when a device is at 100% capacity, generating a queue waiting to be discarded after a particular time if it is a long time, the difference with bottlenecks is that it can be a general problem and not just a device. Bugs in the software of network devices: When the software of the devices has errors. Other: interference or "noise" in the wireless network, proximity to other wireless devices, distance, physical elements (walls), etc. Then, the packet loss leads to the following problems: Information out of date (delay in the information received), slow loading, interruption of loads, closure of connections, incomplete information.
