Quality of Service (QoS) is a set of tools and techniques used in IP networks to manage and optimize the distribution of resources according to the varying needs of different applications, users and services. In this way, it is possible to ensure that applications that require special conditions (such as voice and video) obtain them, even in congestion scenarios (i.e., when the infrastructure is overloaded due to high traffic demand).
This has become increasingly necessary because IP networks were originally designed to provide a best-effort service, which means that all packet traffic is treated equally and sent in a FIFO fashion: first-in, first-out.
What does quality of service consist of?
Quality of service gives the network the ability to provide adequate performance and availability to each application or service running over it, while ensuring certain quality criteria. For example:
- Bandwidth: the amount of data that a network can transmit in a given period of time.
- Latency: time it takes for a packet to traverse the path from its source to its destination.
- Jitter: the variation in latency.
- Loss of packages.
These quality criteria are important, as each application has different requirements.
For example, an application such as Zoom or Teams (Microsoft) demands that the network provides sufficient bandwidth to transmit high quality video and voice in real time, low latency for a better user experience and no image or audio delays, and low packet loss to avoid experiencing interruptions in transmission. In contrast, an email application is not as dependent on latency, as long as it has low loss and sufficient bandwidth to send messages.
How can it be implemented?
There are several mechanisms to differentiate the treatment given to the traffic generated according to your needs:
Traffic classification/marking
Although this does not define the quality of service itself, it is an important part of its implementation, since it allows the network to identify packets as belonging to a certain type of traffic and thus ensure the most appropriate treatment for them.
This can be done in different ways, depending on the characteristics of the network. For example, using the DSCP code in IP packets or the IEEE 802.1p standard to implement it at layer 2 level.
Bandwidth management
Different types of traffic have different bandwidth requirements. Once traffic is identified, based on a marking, as belonging to a certain application or of a certain type, that traffic can be assured of a bandwidth that meets its requirements.
For example, high-quality video and audio traffic should be guaranteed more bandwidth than traffic from e-mail or instant messaging applications.
Congestion behavior
This is fundamental, since the bandwidth of a network path is limited (thinking of cost optimization or investment). By prioritizing sensitive traffic over lower-priority traffic, the smooth operation of important applications is ensured, even in congestion scenarios.
Assured" and "peak" bandwidths can be set for different types of traffic to keep the network operational, even when it is at capacity.
Policing and shaping
These are technologies that allow network traffic to be controlled. Policing is used to limit network traffic to a given level of bandwidth (or packets per second), while shaping makes it possible to impose restrictions on network traffic so that it does not exceed fixed limits; for example, in the case of providing differentiated services to customers contracting different plans.
The use of these tools in all their possible combinations constitutes the implementation of the quality of service, which will depend on the specific objectives and needs of each network.
A technique that adds value to network services
At a time when networks are becoming increasingly important, their ability to dynamically adapt to the needs of different applications with the shortest possible response time is a valuable feature.
The market is relentless and users are becoming more demanding, so designing and implementing an appropriate QoS policy to ensure reliability and optimal infrastructure performance can make the difference between a bad service and a better user experience.
By
Andrés Burel, Engineering Leader of Telco & Smart Cities .
Andrés has a degree in Telecommunications Engineering from ORT University (Uruguay). He has more than 10 years of experience in telecommunications and has worked for vendors and service providers.
