When you have business-critical systems that rely on an internet connection or your business network, it’s not enough to buy expensive gear and throw a lot of resources at the problem. If you want reliable, effective performance, you have to optimize your resources.
Let’s start with an overview look at QoS and what it’s really trying to do. At a basic level, all of the technology you can use to manage traffic and network performance by giving priorities to some traffic types over others constitutes QoS. It’s the network controllers that allow you to utilize this concept.
But, what is the concept?
It’s also pretty simple. QoS tries to maximize traffic flow. In order to do this, it breaks all of your traffic into different categories (or service types). QoS controllers can then turn priorities up or down for different categories so that you get the best performance on the things you care about the most.
This can have downstream impacts on network access, reliability, and even security.
How QoS Works
How does QoS actually optimize things? The core concept falls back on that idea of traffic categories. In fact, QoS allows you to apply identifying labels to all data packets. The identifier is what marks your different service categories, and the QoS optimizations are built around those identifiers.
Here’s another way to look at it. QoS and its value depend entirely on how you define your services. As an example, you can group things by the type of traffic. All of your VOIP applications make up one category. Your videoconferencing comprises another. Your video on demand could be a third. You can keep going through those categories. When they’re identified, your QoS tech will tag each packet accordingly and adjust network resources to give the category you value most the best performance.
While that covers the admin’s point of view, QoS technology looks at things a little differently. Specifically, your QoS controllers focus on performance metrics, and those metrics fit into four ideas: bandwidth, delay, loss, and jitter. These are metrics that generate hard data points, so your QoS maintains the ability to make precise adjustments. That’s where you get the opportunity for advanced optimization.
Let’s consider a simplified example. If your network prioritizes videoconferencing above every other service, then the QoS knows that (because the admin told it as much). The QoS will reduce resources for other types of traffic, enabling videoconferencing to run faster and smoother. The goal is to optimize each of the four metrics to get the best possible performance out of your video conferencing applications.
Now, that’s a simplified example. In practice, you don’t have to narrow the focus of your QoS like that. Instead, you can have rather complicated optimization schemes that prioritize different metrics for different traffic categories. You can even create conditional hierarchies that shift the priorities according to other conditions. This allows the QoS to optimize traffic dynamically, improving specific performances and average performance across the whole network.
How to Best Use QoS
QoS is probably sounding like a powerful tool at this point, but how do you actually use it? How do you know which priorities to set in order to get the best results?
Ultimately, it’s an optimization problem — meaning it’s down to you to design the right hierarchies in order for your QoS to thrive. How do you do that?
The perfect answer varies from one network to the next, but a few best practices can help you think about your specific situation and get you started in a good position.
Identify Your Services
First up, you have to figure out what your service categories look like. You don’t need to reinvent this from the ground up. You can look at common classifications and see if they make sense for your network. Customize the service categories from there as necessary.
Create a Metric Hierarchy
Once you know your service categories, you have to optimize the performance metrics (bandwidth, loss, delay, and jitter). You want to create a hierarchy of which metrics matter the most for each service category. Keep in mind that the most important metrics for video conferencing might not be the same as the most important metrics for a closed-circuit video camera system.
Let’s emphasize that you have a metric hierarchy for each category, and each of these hierarchies is considered independently at this point.
Prioritize Your Services
From here, you have enough information to make qualitative decisions. Rank your services and metrics to flesh out your general QoS guidelines. Make sure that business-critical applications are given the priority they deserve, and allow room for your QoS tools to help you optimize the finer details. As long as you understand where you need to focus your resources in general, QoS will get you there.
Additional Learning Center Resources