Table of Contents
- What is network performance monitoring?
- Network analysis techniques and metrics
- Basic Network Performance Metrics
- Network performance monitoring tools
- Network troubleshooting strategy
- Monitoring network performance with Hosted Graphite
Networks today span the world and provide many connections between geographically disparate data centers, public and private clouds. This creates a variety of network management problems. If your network is not working properly, it can be very difficult or even impossible to get the most productive or correct operation of your applications.
Computer networks can run into problems due to poor software integration, bad connections, insufficient bandwidth, outdated software, and external attacks. These problems reduce the efficiency of your entire infrastructure and can lead to serious negative financial consequences.
A sophisticated network requires constant monitoring using the right tools and creating a network performance monitoring strategy. One of these tools is MetricFire, a powerful SaaS product that specializes in monitoring systems. Among other things, it offers a comprehensive network performance monitoring solution that will be optimized by our specialists specifically for your purposes.
What is network performance monitoring?
Network performance monitoring is a process of collecting and analyzing data about network traffic that flows across your IT environment. The main purpose of network performance monitoring is to obtain the necessary information about the state of the network to take effective management decisions. All network components such as routers, switches, firewalls, servers, and virtual machines are monitored.
Depending on the complexity of the infrastructure and the purpose of monitoring, the data collected may include basic metrics of network performance, for example, bandwidth, throughput, packet loss, etc. as well as more complex packet data. Batch network analysis allows you to see not only indicators of the state of the infrastructure, but also metrics of the application performance in terms of each user operation, session, the response time of databases, application servers, as well as the time of a request through the network, details of a user request and response, etc. This provides an accurate understanding of the impact the IT infrastructure has on the operation of business applications.
By analyzing all the information collected, network performance monitoring allows you to find and investigate network problems that can lead to the unavailability of applications or slowdowns in response. Network monitoring is also important because it helps determine if the root cause of performance issues is in the network, in the application, or in the host infrastructure.
Network analysis techniques and metrics
In order to build a network monitoring system, first, you need to decide which monitoring techniques you will use. There are three main techniques for monitoring network performance:
- Flow monitoring: this method collects and analyzes data streams generated by network equipment and presents them in a user-friendly format. With this technique, you can understand which devices are communicating, how long they are communicating, and how often. It should also give information about how much data is being transferred. This allows you to identify devices that are degrading bandwidth, find bottlenecks in your system, and improve overall network efficiency.
- SNMP (Simple Network Management Protocol) monitoring: SNMP is a common protocol, the task of which is to manage network devices and obtain information about their work. As a rule, all modern managed network devices (workstations, laptops, switches, printers, routers, modems, webcams, etc.) that support SNMP have a so-called Management Information Base (MIB). This database contains a lot of useful information about the state of the device: performance counters, active processes, network traffic values, etc. SNMP monitoring allows you to discover endpoints on your network and analyze traffic flowing to them. The main limitation is that SNMP monitoring does not perform well in public clouds or other environments where there is no access to network equipment. Sometimes it is possible to monitor SNMP in the cloud by installing so-called SNMP traps.
- Packet capturing is the process of capturing and registering traffic. As data streams pass over the network, the analyzer captures each packet and, if necessary, decodes the raw data of the packet, showing the values of various fields in the packet, and analyzes its contents. Packet capture aids in monitoring network health as well as application network performance, monitoring security and incident responses, troubleshooting service issues, and network capacity planning.
Basic Network Performance Metrics
Regardless of which monitoring method(s) you use, most likely you will analyze the following basic network performance metrics:
- Throughput: a measure of how many units of information a system can process in a given period of time. It is more of a speed indicator as it relates to aspects such as response time.
- Bandwidth: a measure of the maximum amount of data that can be transferred over a network in a given period of time. Typically this time period is measured in bits per second. Bandwidth is a measure of capacity, not speed.
- Uptime: the total time the network remains up and available.
- Latency: a measure of how long it takes for a data packet to travel from one designated point to another.
- Packet loss: shows how many packets fail to reach their destination.
You can start monitoring with these key metrics, gradually expanding them depending on your needs and the specifics of your infrastructure.
If you would like to dive deeper into network performance metrics, we have a post on this topic.
Network performance monitoring tools
Network monitoring solutions can be roughly divided into three classes.
Open source tools
The first is various open-source tools that can be downloaded and used for free. But, like many free products, they are not supplied in a "box" form and require fine-tuning for specific tasks, which requires the presence of qualified specialists on the staff. In this case, all responsibility for the operation of the monitoring system lies with the specialist. The company should take into account the fact that a specialist may quit and it will be very difficult to understand their settings. The use of open-source programs is quite justified when solving basic, non-trivial monitoring tasks.
Your network provider’s built-in monitoring software
The second class of solutions is the monitoring tools included in the products of other manufacturers. For example, companies-suppliers of virtualization tools, as well as network infrastructure equipment, offer ready-made monitoring systems for their solutions. These professional tools are based on the world's best practices. Manufacturers are responsible for their development and support. But you need to understand that the functionality of such a solution can be limited to working only with a certain set of equipment or systems.
3rd party proprietary monitoring solution
The third class is 3rd party proprietary monitoring solutions monitoring solutions. Their developers are focused and specialized in developing products for in-depth analysis of the performance of network infrastructure and offer the most functional solutions on the market. Enterprise-level monitoring tools offer not just an analysis of the state of the network in terms of its speed or latency, they are tools for monitoring the quality of business applications from the point of view of the network interaction of its participants.
MetricFire belongs to the third class of tools and at the same time is based on open-source projects. Please book a demo, or sign up for the free trial to display your network performance metrics on highly functional Grafana dashboards using hosted Prometheus and Graphite solutions.
Network troubleshooting strategy
In order to avoid the critical impact of network problems on your users, you need to develop an effective network troubleshooting strategy. The result of the strategy will be the systematization of the process of determining the nature of the problem, the classification of incidents, and the response to it according to a predetermined algorithm of actions.
The key to a good strategy is detailed documentation. Write down the steps you will take to assess and diagnose problems, and the response plan you intend to follow. Having this documentation will help you respond more quickly if a problem occurs. Update the documentation regularly based on your experience.
Let's take a look at the basic steps for resolving network problems.
1. Find the nature of the problem
Before you start solving a problem, you need to determine what exactly caused it. Often, user complaints are vague and can mean anything. It is important to find what was the reason for contacting you. Monitoring network performance metrics over time will greatly simplify this task.
2. Classify the problem
Once the nature of the problem has become clear, you need to assign a priority level to it. The classification of incidents reflects the scale of the problem - how many users are affected by the incident, how much it interferes with their work. Depending on these indicators, each level of incident should be assigned a schedule for resolving the problem.
3. Develop an emergency plan
To effectively deal with network performance issues, you must first approve an emergency plan. This plan should contain procedures and resources to address specific network problems. It can also include indicators that can signal potential problems and ways to solve them before they affect users. An emergency plan needs to be tested and reviewed regularly so that it is as close to real conditions as possible.
4. Add the incident to the database
Each incident must be recorded in the database. This database should at least contain the date/time of the incident, its description, its classification, the actions taken to eliminate it and the resources spent. Maintaining a database of incidents will make it possible to more accurately develop an emergency plan, as well as observe the frequency of certain problems. Analysis of the incident database can reveal both equipment problems and repetitive human operational errors.
Monitoring network performance with Hosted Graphite
MetricFire’s Hosted Graphite is based on the open-source monitoring tool Graphite that provides powerful network performance monitoring capabilities. It provides a collection of historical network data from certain metrics that can be used to build statistics for management or clients and identify behavioral anomalies that could indicate problems. Since Hosted Graphite is based on open-source solutions, it is highly customizable and flexible in the setting. MetricFire’s monitoring solutions support companies of all sizes.
MetricFire’s Hosted Graphite goes with the Grafana visualization tool that makes it real to view your network monitoring data on beautiful informative dashboards in real-time. Another feature of MetricFire’s monitoring solution is alerted. Alerts are a monitoring element that takes action based on changes in metric values. Alerts keep users informed of important events, even if they are not physically present to monitor metrics in their dashboard.
Hosted Graphite is also designed to be completely cloud-based. This means that it does not require any software to be installed on the user's local hosts. To learn more about monitoring network performance with MetricFire’s Hosted Graphite, check out our article.
Your network needs to be managed and monitored just like the rest of your IT infrastructure. Correctly defined metrics, their critical values, a good troubleshooting plan, and usage of the right monitoring tools is the key to a successful network performance monitoring strategy.
MetricFire has a cloud-based network monitoring solution that perfectly matches your scale. Usage of our product requires minimal configuration to gain in-depth insight into your environments. Our team of professionals can recommend a network monitoring solution that will work best in your situation.