Using Qosium is typically either manual measuring or automatic monitoring. In monitoring, Qosium observes, for example, that a critical service continuously gets the QoS it needs from the network. Monitoring can range from a single link and application to the whole network.
Toolbox
When monitoring, the following components are typically selected in the toolbox:
Measurement agents: Qosium Probe
Measurement controllers: Qosium Scopemon
Results system: Qosium Storage or customer’s own system
Like always with Qosium, whether it is about manual measuring or automatic monitoring, Qosium Probes handle almost all the measurement-related calculations, so they need to locate in places where they see the traffic to be monitored. Qosium Probe is typically run in the background as a system service.
Automatic Control & Gathering Results
The usage determines the selection of the measurement controller. When measuring, Qosium Scope is selected, but when monitoring, the measurement is controlled by Qosium Scopemon. It is parameterized and turned on, after which it automatically does its job. Sometimes the users need to see a high-level overview of their network quality, making Scopemon’s straightforward user interface handy. It can also be parameterized to perform simple alarms. In many use cases, however, the user interface is unnecessary, so Scopemon can run in the background performing its tasks hidden from the user. The document section talks more about Scopemon.
The results system, where all the measurements send their results continuously, has a significant role in monitoring. By using Qosium Storage, you can see the status of the whole network with a glance.
The results system can also be your own. There are many ways to integrate Qosium results reception directly into your system.
Using a results system is not mandatory, but it glues the pieces of the monitoring as an entity. If the monitoring setup is small and composed only of a few measurements, dealing with result files might be adequate.
Monitoring Scenarios
The measurement agents are installed in places that are of importance concerning monitoring. They can be seeded in the network to be ready for upcoming possible monitoring. When a measurement is not active, the resource usage of Qosium Probe is negligible. Regarding the placement of the agents, there are two things to consider:
Qosium Probes must see the monitored network traffic and
QoS monitoring is focused on the network path between two Probes.
If monitoring is carried out at a single point, the available results set becomes limited, and QoS cannot be measured with a few exceptions. In the setup, there will be an instance of the measurement controller per single QoS measurement. In the optimal situation, the measurement controller is in the same device as the Probe it uses in monitoring, but it can also be elsewhere.
A single results system can typically handle the whole monitoring entity. In extensive and some special setups, multiple results systems might be needed. You might also want to consider duplicating the results system in critical environments
An exemplary monitoring scenario is depicted below. The connections between workstations and application servers are monitored. Qosium Scopemon is parameterized to automatically wake up a two-point QoS measurement when a workstation starts using a monitored service. The results are delivered to Qosium Storage, located in the same network, and Qosium Storage shows the status of the whole network.
In the example above, the connections are monitored end-to-end, but that’s not always the case. The scenario could also monitor only a particular part of the network or even single links. Qosium Probe is the one whose location defines the monitoring and measurement possibilities. Thus, it can be installed in various environments on almost all types of desktop and server machines, lightweight end devices, and many embedded systems. We continuously extend Qosium to new platforms and operating systems based on the need.
The document section talks more about Qosium Probe.
Situation Awareness
The information provided by monitoring is observed using a results system like Qosium Storage.
Instead of Qosium Storage, you can use your system that takes in and manipulates as needed the results created by Qosium. In many use cases, the monitoring results provided by Qosium are enough as such for situation awareness, while in some cases, they are just a piece of a vast set of results generated by different tools.
Qosium Storage shows the status of the monitored network as a whole. It is possible to create reports, for example, daily, telling the desired information of the monitored system. In Qosium Storage’s QoS heatmap visualization, the situation awareness is in its most illustrative form. It resembles that of Qosium Scope and is, in fact, based partially on the same technology. The main difference is that while Qosium Scope visualizes a single measurement, Qosium Storage’s heatmap is able to visualize the monitored system as a whole.
You can select the measurements or even all of them to be visualized. There are many criteria for the selections, like time scale and different measurement identifications. The identifications can be mapped, for example, to single network devices, enabling per-device QoS inspection.
Areas of weak QoS are found instantly, and deviant statistics can be pinpointed. It is easy to check which device got bad service from the network, where, and when. After that, you can focus the inspection. If the problem is observed only by a single device, it is likely that the problem is somehow related to that device. Instead, if other devices also get weak QoS in the same area, it is likely a problem on the network side. By going deeper, you can see is it about a single base station or a cell. Qosium’s physical layer statistics tell already a lot that is the root cause of the problem to be found in the radio access part or on the network side.
Detailed Analysis and External Tools
In addition to visualizations and reports, Qosium Storage gives general information about the system status and lists all the stored measurements. You can browse and search results, and when something interesting is round, access the raw results. Connecting results to external tools is also easy.
The raw results can be examined with Qosium Storage, but often it is feasible to do, for example, the more accurate statistical analysis with other tools. That is straightforward since you can download the raw results of individual measurements directly to your local workstation. The results file format is the same CSVComma-Separated Values Originally CSV was only about values that were separated by commas, but nowadays the separation mark can also be something else like a semicolon. format all the other Qosium components use. Thus, you can, for instance, feed the downloaded results file directly to Qosium Scope, enabling accurate spatial and temporal analysis. However, nothing stops you from using your favorite external tool (MATLAB, Excel, R, Jupyter, etc.).
The database of Qosium Storage is PostgreSQL by default. Therefore, your system can also read Storage’s database directly, connecting Qosium statistics as part of a seamless network management experience. Qosium Storage uses the database in timescale format, enabling the results to be directly visualized by popular tools like Grafana.
Taking Qosium monitoring into use raises network situation awareness to a new level. The crucial difference compared to traditional monitoring is that while it tells sporadic things about network devices and perhaps the quality of artificial test traffic streams, it does not reflect well the real quality of applications and services from the end-user perspective. Qosium tells exactly how users, whether they are humans or machines, see the network quality.
