Article Preview
TopIntroduction
Over the past two decades, we have witnessed a growth in the number of interconnected devices with the integration of sensors and smart objects to the internet. A new term has emerged in the Internet vocabulary known as IoT “Internet of Things (Goyal, Sahoo, Sharma, & Singh, 2021). The IoT represents the next evolution of the internet. It reflects an environment by which billions of smart products around us in various fields (industrial, healthcare, smart cities…) are networked and connected to the internet. According to a new analysis from IHS Markit (Nasdaq, 2017), a world leader in critical information, analytics, and solutions, the number of connected IoT devices worldwide will jump 12% on average annually, from nearly 27 billion in 2017 to 125 billion in 2030.
M2M (Machine to Machine) is one of the main pillars of IoT. It is a communication paradigm where devices are mostly unsupervised (Kim, Lee, Kim, & Yun, 2014). It provides ubiquitous connectivity between devices with the ability to communicate autonomously without human intervention. Moreover, devices can make autonomous decisions based on information received from other devices, making possible new opportunities for multiple applications and bringing various benefits to our lives. IoT and M2M communications are expected to be a major communication paradigm in the future Internet (Prasad & Rohokale, 2020).
Depending on the context where the application is invoked, Quality of Service (QoS) requirements can be determined. Generally, QoS refers to any technology that manages data traffic and network resources to reduce packet loss, throughput, latency, and jitter by placing priorities for particular data on the network.
IoT and M2M applications require different QoS support (availability, reliability, scalability, interoperability, and security) to provide real-time, duplicate-free information, and make appropriate decisions without any errors (Kesavan & Prabhu, 2018). In addition, in these systems, message loss and delays are unacceptable and can cause system instability, what could lead to physical damage and a risk to security and human safety.
In this context, several contributions have emerged in the literature. A well-known application is WSAN (Wireless Sensor Actuators Network), deployed to monitor and react against forest fires (Kułakowski, Calle, & Marzo, 2013). In this scenario, sensors gather temperature data from the environment and then locate the fire areas, and finally, the actuators will react by dropping water on fire areas.
The actuators are machines, and their actions depend on the collected data by sensors, so they can be informed where and when they should intervene. Thus, it is crucial for sensors to deliver data with high QoS in terms of reliability and less delay. The contrary can cause system instability and lead to human and material damage.
In such a scenario, where various types of messages are used (simple info messages, alert messages, and critical status messages), the application layer protocols need to meet QoS requirements for each type of message. For example, a normal temperature value gathered by sensors should not be delivered with high reliability since it does not require any action by actuators. However, suppose the temperature is too high, which means there is an incident; In this case, critical messages should be transmitted to actuators with reliability and in time-limited bound to react rightly in real-time. Thus, it is critical to provide reliable delivery of messages, with less time to avoid any disaster damages.