When speaking with customers about what they need to support their infrastructure, the thing we hear most is “network performance monitoring.” This blog will share my insight on what is network performance monitoring, tools utilized, and methods to use Scrutinizer for monitoring network performance.
What is network performance monitoring?
Network performance monitoring means to measure, evaluate, and optimize the integrity of flow sources such as cloud, data centers, SD-WAN, virtual machines, applications, and even users. Network performance monitoring provides a proactive approach to resolve issues quickly to reduce downtime and costs.
Once network benchmarks are set, IT Ops builds a capacity plan and gathers network traffic measurements such as throughput, usage, packet loss, and jitter. Using these network performance metrics allows IT Ops to build QoS policies and traffic shaping for the network infrastructure.
As legacy infrastructures transition to robust architectures such as the cloud, software-defined networks, or virtualized platforms, IT Ops used a variety of solutions to benchmark the network.
There is Ipswitch’s WhatsUp Gold solution, which uses NetFlow to monitor network traffic and bandwidth usage from flow-capable devices. But WhatsUp Gold has a limitation of device roles and discovery, making it a challenge to obtain the full conversation from network devices without adding additional modules.
SolarWinds Orion is another solution that performs network performance monitoring, but requires the user to deploy an agent to do so. Using this functionality leads to inconsistent and unreliable auto-discovery, which does not grasp the full conversation of the traffic from those devices.
Paessler’s PRTG is good with monitoring bandwidth usage and uptime, as well as collecting metrics from network entities and applications. PRTG has an auto-discovery mechanism that scans predefined areas of an enterprise network and creates a device list from this data. PRTG lacks real-time monitoring and customize reporting, requires sensors (which loses visibility), and has limitations on alerting capability.
Scrutinizer provides support of any flow format without the use of agents or upgrades. With it, IT Ops can build a workflow and monitor an application transitioning to the cloud network. Scrutinizer allows a user to drill into a single host, traffic pattern, or violator to detect the bandwidth hog. Scrutinizer provides contextual visibility, auto-discovery, incident management, traffic analysis, threat forensics, and security analytics. It can also integrate with SIEM platforms and configuration management databases (CMDB).
Best practices to troubleshoot network performance issues
There are some best practices when evaluating network performance issues. To start, ask these questions when troubleshooting poor network performance:
- What network performance monitoring tools are used to detect the problem?
- Are there any abnormalities in the network?
- Is there a packet capture?
- Who is the top talker consuming bandwidth?
Let’s take, for example, where NetOps experiences a soft phone issue. When using a network performance monitoring solution, the first step of investigation is to verify if there is packet loss, latency, and/or retransmission in the network.
The next step is to investigate the application latency, client and server uptime and downtime, and the quality of service (QoS) priority assigned to the soft phone. From a NetOps perspective, VoIP requires high priority. VoIP metrics such DSCP, CBQoS, RTP payload, and jitter are key performance indicators to evaluate the network-performance.
The final aspect of investigation is the end user activity and understanding what resources are consuming the network, which affects his/her VoIP quality. Having a complete solution to monitor these metrics is important in resolving the issue swiftly.
Let’s take a look at this report; using Scrutinizer gives you a flow analysis of these VoIP metrics into the network.
Methods to monitor network performance
There are several network performance monitoring solutions that use different methods to monitor the performance of the network. One option is SNMP, which manages the CPU utilization, bandwidth utilization of links, uptime, memory, and disk usage of network devices. Another option is Windows Management Instrumentation (WMI), which focuses heavily on performance monitoring of Windows devices. Using the NetFlow and IPFIX implementations supports auto-discovery, remote
visibility, predictive analysis, threat forensics, and network mapping from any flow source device. Other methods to detect device behavior are ping, HTTP, and RESTful API.
Network performance monitoring needs to be customized to view the entire network infrastructure. Scrutinizer provides users with:
- Customizable dashboards
- Visibility of all vendors’ metadata using unique Netflow and IPFIX elements
- Advanced reporting and filtering
- The ability to integrate with other platforms and have alerting methods such as syslog, email, SNMP, and RESTful APIs.
Network performance monitoring solutions should have the ability to monitor a wide array of vendors and help the NetOps team with forecasting and capacity planning. That said, configure NetFlow for your devices and download Scrutinizer, and become network-aware.