Network Performance Management Demystified: Key Metrics and Modern Solutions

Network performance management mainly focuses on obtaining some helpful insights regarding the operation of the network and determining potential risks. Here, in the article, we will dive deeper into all the key metrics and modern solutions in network performance management.

Why Effective Network Performance Management Matters

Network performance management is extremely important for getting all the valuable insights regarding the network’s health and offering some helpful improvements. It includes all the needed activities to maintain a better performance.

Initially, it includes the gathering of all the crucial network aspects for the evaluation of the current performance. With the help of gathering information, it becomes way easier to detect some troubles, optimize performance, analyze some insights, and even create a better capacity planning.

The major role of performance management lies in the prevention of serious troubles and consequences. That’s why you can create a proactive environment with a higher reliability and availability rate.   

Core Components of Managing Network Performance

Here are the major processes that you should definitely know about:

• Identification of the performance troubles can help with an immediate fix of the issues without any serious consequences.
• Reporting and forecasting can help with the prediction of future tendencies based on deeper analysis.
• Daily modifications of the operational performance directly impact the changes in accommodation in the traffic layout.
• Capacity planning functions on the basis of historical data and can help with future capacity expansion.

Leveraging Performance Metrics for Insights

For the proper gathering of the performance metrics, all the information should be properly stored, processed, and analyzed. Any kind of monitoring system should offer quick access to all basic metrics on all devices. Usually, information should be categorized based on the time frame, together with the historical data and a combination of various metrics.

In the statistical parameters of the performance management, the information is collected in counter and gauge metrics. A gauge represents a specific numeric value that can change depending on the circumstances, such as CPU utilization or current usage of memory. The monitoring of CPU is the most standard example of a gauge, as this number can increase or decrease based on its utilization.

A counter is the cumulative metric that only increases or can be reset to zero. Such metrics can relate to the number of processed requests or the number of total errors. This number can only increase with time, and there won’t be a sudden drop in the already conducted requests.

There might be some exceptions to the standard measurement. For instance, when monitoring with SNMP, counters are used to monitor the number of sent/received packets and bytes. These numbers increase and have a cumulative total that can be extremely helpful for the trendcy detection. Of course, the interface traffic can fluctuate, but counters help to measure total traffic volume of data.

Performance information is in specific contexts, so the performance management system should offer access to the additional context, especially if something went wrong. To the additional context, we can relate device location, device data, customer, responsible user, and more.

Statistical parameters are such as main traffic hours, average utilization, percentage of customer connections, and more. This information can be helpful in indicating some potential issues and risks.  

Key NPM KPIs to Monitor

KPIs are specific metrics that are needed for the evaluation of the efficiency and effectiveness of the network. It can help IT experts to monitor such aspects of performance as bandwidth, packet loss, latency, and more. By tracking these metrics, the company can be sure of the optimal functioning of the network, determine possible future risks, and make more informed decisions regarding some improvements.

KPIs can be obtained either from the calculation of several metrics or directly from the network. There are lots of various KPIs to be tracked as well, and new ones can be created based on the requirements of the project.

When talking about some KPIs in the mobile networks, they can be the following: RSRQ, RSRP, SINR, and many others.  

Identifying and Addressing Network Slowdowns

Another important process in the performance management relates to the slowdowns and other possible anomalies. This may be done with the use of dynamic or fixed thresholds. When the metric value becomes lower/higher than the threshold, then there will be an alarm notifying about it.

Thresholds may be set at a fixed percentage value for a specific time, or even a dynamic approach can be used. The threshold can be determined as a deviation from the baseline. This baseline can be determined with the help of historical metrics.

Triggering Automated Responses

Once a specific degradation is determined, usually the alarm is triggered. With such an automated response, IT professionals notice that something is not going according to plan. There are various automated responses to some changes.

For example, a critical event in any data center is a sudden temperature spike that should be immediately notified of. The system should immediately notify experts or even react with immediate actions. By triggering the HVAC system, urgent cooling can start.

Capacity Forecasting and Resource Planning

We have already discussed that performance management includes all the activities needed for maintaining the infrastructure as close to the optimum as it is possible.

This relates to various short-term or daily routines such as tilting antennas, changing the faulty devices, fixing optical cable attenuation, and performing other activities. Nevertheless, all such activities won’t be a huge help in case the digital infrastructure is insufficient.

Because of this, IT companies and some telecoms usually have strategic planning units that are working on expanding, optimizing, and even modernizing the infrastructure to speed up the daily operations and guarantee optimal service functioning. To achieve the maximum possible result, such activities as capacity planning and capacity management are necessary. Moreover, it is crucial to also consider the requirements of standard operations and business decisions.

Regular attention to the capacity is highly important, and it includes the following important activities:

• Detecting some performance troubles and analyzing performance from a historical perspective.
• Recording the performance of the infrastructure and performance monitoring.
• Optimizing existing capacity to take business decisions into account.
• Implement some changes in the infrastructure to align with the plans.
• Planning some capacity expansion to meet requirements such as the introduction of the new service.

As you see, capacity management includes lots of activities, and there is no chance to share all of them in one post. So, check other details if you need information regarding capacity planning.  

Performance Optimization and Reporting

Networks need almost daily operational performance improvement, specifically ones with dynamic behavior. IT experts make their decisions based on the performance data. To simplify the decision-making process, there should be used flexible reporting that offers access to the relevant data. Here are several examples of the activities and reports:

• SIP truck statistics: gives access to insights regarding SIP truck usage of the upstream telco voice trunk.
• Top N interface utilization: specifies which interface is the most used, so that it is possible to balance traffic and redistribute traffic load.
• Number of connected users for Wi-fi access point: helps with the redistribution of the load.
• Latency Report: determines the time that is needed for the information to go from one point to another.
• Top N temperature for data room: gives access to the information regarding the temperature to make some decisions regarding redistribution of power or temperature load.
• Packet loss report: calculates the percentage of lost packets, which can specify the network-specific troubles.
• Traffic analysis report: divides traffic according to the destination, type, and source.
• Top talkers report: determines the users and devices generating the major traffic, and that information is crucial for bandwidth management.
• Capacity planning report: gives information regarding current resource use and forecasts the future requirements.
• Security incident report: malware detection or unauthorized access attempt helps guarantee network security.

Driving Ongoing Process Enhancements

The more IT experts get insights regarding the network health, the more they can determine the problems in the procedures and processes. This gives a possibility to introduce needed steps for the improvement and impulse to revise the processes continuously.

With the continuous reporting, it is possible to give a solid foundation for the execution of the improvements in the procedures and processes.

How NPM Relates to Network Monitoring and Fault Management

NPM, as well as network monitoring and fault management, are important aspects that should be focused on. When working specifically only on one of these aspects, it is possible to miss something important. Detection of pending performance degradation and immediate reaction guarantees proper and long-term functioning of services.

Let’s discuss one example so you will have a better understanding of how everything functions. In case there is a 1 Gb/s internet uplink, which is used by a team of employees. Then suddenly, a new cloud application that requires a lot of bandwidth is overutilized, and some issues appear. Theoretically, there are no issues with the network, and all the components function as needed.

Only with a correctly configured fault management, is it possible to determine this drop rate and trigger an alarm. That is where performance management is so important to understand the root of the problem.

Also, a good performance management system has thresholds set. For instance, 70 % will show the warning alarm, 80% will show the major alarm, and 90% will show the critical alarm. This will help engineers identify the problem much earlier than it will impact the actual customers.

Best Practices for Network Performance Optimization

The choice of the best way of management approach depends on the architecture, network use, and type. However, there are some generic recommendations for most situations:

• Define KPIs and relevant performance metrics: it is impossible to gather all the performance metrics, which is why it is important to limit the metrics to the most relevant information.
• Real-time consolidated monitoring: to guarantee the best possible performance management, it is highly important to collect information in real time. Data collection relates to the information from the network devices, network elements, data collected by passive and active probes, environmental sensors, and more.
• Alerts and thresholds: start with the determination of the baseline behavior, so the trigger alarm will switch on when the lower or upper metric values are received. A great idea is to specify minor, major, and critical thresholds so you will receive progressive alerts regarding the situation.
• Performance reports: with the usage of N device memory utilization and other similar reports, it will be possible to identify and prevent future faults.
• Regularly update procedures and processes: the best possible practice is to clearly understand the actions that should be taken when something goes wrong with the performance.
• Additional context for performance data: by utilizing additional context data, it is possible to isolate and localize the issue much faster.   

Typical Obstacles in Managing Network Performance

Even after the utilization of the best practices that we have already shared, there might still be some challenges. So, let’s review some of the most common troubles that can occur during the introduction of performance management.

One of the most frequent issues is establishing the connection between the network elements and PMS. Another common problem relates to the scalability. This issue is connected with the technology implemented in PMS.

One more challenge is related to the integration of the monitoring system and the different systems that should be monitored. For example, lots of systems use proprietary protocols that need a better experience in the integration processes.

Consolidating monitoring is always a huge challenge, because you cannot be sure that all the elements are being monitored. To deal with it, you should use automatic solutions.

Emerging Trends Shaping the Future of NPM

• Usage of AI/ML: all the learning algorithms can significantly simplify the processes of capacity planning, forecasting, baselining, and other aspects.
• Umbrella management: 360-view or end-to-end visibility becomes a huge standard for modern management and monitoring.
• Migration to the cloud: there is a huge trend towards cloud migration. Despite some arguments regarding this direction, this is a fantastic option for lots of projects.
• Adoption of the new technologies: among the most popular new trends, we can specify edge computing, 5G, IoT, and others.
• Integrity with the security systems: security level can be significantly boosted with health data and other important information.