AF (Application function/Assured forwarding)

Introduction:

Assured forwarding (AF) is a service model defined by the Internet Engineering Task Force (IETF) for providing Quality of Service (QoS) in IP networks. It is designed to provide predictable and reliable network performance for applications that require a certain level of service quality. AF is one of the four service models defined in the Differentiated Services (DiffServ) architecture. The other three service models are best effort, expedited forwarding, and class selector.

What is AF (Application function/Assured forwarding)?

AF is a QoS service model that provides a predictable level of network performance for traffic that requires a certain level of service quality. It is designed to provide reliable and predictable network performance for applications that require a certain level of service quality. AF is based on a set of predefined traffic classes that are defined by the network operator. Each traffic class is assigned a set of service parameters that define the level of service quality that will be provided for traffic in that class.

AF works by assigning traffic to one of four classes (AF1, AF2, AF3, or AF4), each of which is assigned a different level of service quality. The four AF classes are defined as follows:

  • AF1: This class provides a low level of service quality and is suitable for traffic that can tolerate some loss and delay. Examples of traffic that can be assigned to this class include non-critical data applications such as email or file transfer.
  • AF2: This class provides a higher level of service quality than AF1 and is suitable for traffic that requires a moderate level of service quality. Examples of traffic that can be assigned to this class include voice over IP (VoIP) and video conferencing.
  • AF3: This class provides a higher level of service quality than AF2 and is suitable for traffic that requires a high level of service quality. Examples of traffic that can be assigned to this class include mission-critical data applications such as financial transactions or healthcare data.
  • AF4: This class provides the highest level of service quality and is suitable for traffic that requires the highest level of service quality. Examples of traffic that can be assigned to this class include real-time applications such as video streaming or online gaming.

Each traffic class is assigned a set of service parameters that define the level of service quality that will be provided for traffic in that class. These service parameters include packet loss, delay, and jitter. Packet loss refers to the percentage of packets that are lost during transmission, delay refers to the amount of time it takes for a packet to travel from the source to the destination, and jitter refers to the variation in delay between packets.

AF uses a combination of traffic classification and traffic conditioning to provide the desired level of service quality for each traffic class. Traffic classification is the process of identifying the traffic class to which a packet belongs based on its source address, destination address, protocol type, and other header fields. Traffic conditioning is the process of shaping or policing the traffic in a particular traffic class to ensure that it conforms to the service parameters assigned to that class.

Benefits of AF:

AF provides several benefits for network operators and end-users. These benefits include:

  • Predictable network performance: AF provides a predictable level of network performance for traffic that requires a certain level of service quality. This can be especially important for applications that require real-time performance, such as VoIP or video conferencing.
  • Efficient use of network resources: AF enables network operators to allocate network resources more efficiently by assigning different levels of service quality to different traffic classes. This can help to reduce congestion and improve overall network performance.
  • Improved end-user experience: AF can improve the end-user experience by ensuring that applications that require a certain level of service quality receive the appropriate level of service quality.
  • Flexible service configuration: AF provides network operators with a high degree of flexibility in configuring QoS policies. They can assign different service parameters to different traffic classes, allowing them to prioritize certain types of traffic over others.
  • Scalability: AF is a scalable solution that can be implemented in large-scale networks without incurring significant overhead or complexity. It can be deployed in both wired and wireless networks, making it a versatile solution for a wide range of network environments.

AF Implementation:

AF is implemented using the Differentiated Services (DiffServ) architecture, which is a QoS mechanism that operates at the network layer (Layer 3) of the OSI model. DiffServ uses a set of code points (DSCP) to mark packets with the appropriate service class. These code points are defined in the IP header and are used by routers and switches to identify the appropriate service class for each packet.

The implementation of AF involves two key components: traffic classification and traffic conditioning. Traffic classification involves identifying the traffic class to which a packet belongs, based on its source address, destination address, protocol type, and other header fields. Traffic conditioning involves shaping or policing the traffic in a particular traffic class to ensure that it conforms to the service parameters assigned to that class.

Traffic Classification:

Traffic classification is the process of identifying the traffic class to which a packet belongs. This is typically done by examining the packet header and comparing it to a set of predefined rules that are configured on the network device. The rules are defined based on the source and destination IP addresses, protocol type, and other header fields.

Traffic conditioning:

Traffic conditioning is the process of shaping or policing the traffic in a particular traffic class to ensure that it conforms to the service parameters assigned to that class. This is typically done using queuing and scheduling mechanisms. Queuing mechanisms are used to manage the flow of traffic in a particular traffic class, while scheduling mechanisms are used to determine which packets are sent first.

In AF, traffic conditioning is typically done using the following mechanisms:

  • Traffic shaping: Traffic shaping is used to limit the rate of traffic in a particular traffic class to a specific bandwidth. This helps to prevent congestion and ensure that the traffic in the class conforms to the service parameters assigned to that class.
  • Traffic policing: Traffic policing is used to enforce the service parameters assigned to a particular traffic class. If a packet exceeds the service parameters, it is either dropped or marked with a lower priority code point.
  • Queuing: Queuing mechanisms are used to manage the flow of traffic in a particular traffic class. Different queuing mechanisms are used for each traffic class, depending on the level of service quality required. For example, a low-latency queuing mechanism may be used for real-time traffic, while a weighted fair queuing mechanism may be used for non-real-time traffic.

Conclusion:

Assured forwarding (AF) is a QoS service model that provides a predictable and reliable level of network performance for traffic that requires a certain level of service quality. It is designed to provide predictable and reliable network performance for applications that require a certain level of service quality. AF is based on a set of predefined traffic classes that are assigned a set of service parameters that define the level of service quality that will be provided for traffic in that class. AF provides several benefits for network operators and end-users, including predictable network performance, efficient use of network resources, improved end-user experience, flexible service configuration, and scalability. The implementation of AF involves traffic classification and traffic conditioning, which are used to identify the traffic class to which a packet belongs and shape or police the traffic in that class to ensure that it conforms to the service parameters assigned to that class.