QoS Models

For a lot of people QoS is a very hard concept to grasp and understand. This is understandable as there is a lot that goes into it and a lot to learn. In this post I want to go over the different types of QoS Models. There are three main models that are used for providing QoS in networks. These three are: Best Effort, Integrated Services (IntServ), and lastly Differentiated Services (DiffServ). So what exactly are these and what do they mean. I want to start off by explaining the easiest to understand which is Best Effort.

Out of the three models, the Best Effort (BE) model is the simplest. It is the default QoS model used on the internet and does not apply any QoS mechanisms. Traffic is delivered on a first-come, first-serve basis, and no special treatment is given to any type of traffic. While this model doesn't perform well in environments with real-time traffic demands, it can still provide a reliable level of service. The BE model offers maximum network efficiency and cost-effectiveness, depending on the needs of the network.

The Integrated Services (IntServ) model allows us to signal and reserve network resources for specific applications using the Resource Reservation Protocol (RSVP). However, it's not ideal for large-scale networks, as every node along the path must support and maintain RSVP state information for each individual flow to ensure end-to-end QoS.

The Differentiated Services (DiffServ) model enables traffic prioritization based on predefined criteria, such as data type and origin. It uses techniques like traffic shaping where excess traffic is buffered and traffic policing, which drops packets that exceed a specified bitrate. One major advantage of DiffServ is its scalability, as it can handle many traffic classes. However, it relies heavily on correct implementation and configuration. It is commonly used in enterprises.