Network Monitoring Tools

How to Identify and Troubleshoot Packet Loss

What is Packet Loss?

During data transmission, the information is split into smaller sized packets, and sent across the network or internet. However, sometimes not all the packets make it to the destination, and results in slow service or network errors. This is called packet loss and can be extremely disruptive to users accessing company resources or the internet. Packet loss is a problem that can affect any network but is especially present in Wide Area Networks.

Packet loss can be caused by multiple factors:

  • Faulty hardware or configuration
  • Link failure, poor network performance
  • Network congestion
  • Buffer overflows
  • Misrouted packets
  • Cyber-attacks

 

Problems Caused by Packet Loss

The effects of Packet Loss vary based on what transmission protocol is used. For instance, in TCP where a connection is established before data transfer, packets that are lost are usually retransmitted. However, this can lead to a reduction of speed, severely affect the throughput of a given connection and generates extra traffic because the data may need to be retransmitted. These delays cause high latency, which means a slow delivery time. On the other hand, communications that use UDP such as VoIP, video conferencing, and video streaming do not have built-in retransmission capability and sacrifices data integrity in exchange for speed. As a result, it cannot properly handle placket loss and when it occurs it leads to a loss of data, or poor quality of service. Impairments such as short gaps or robotic quirks in VoIP conversations, pauses, jumps, and pixelated frames in live video streams are the result of packet loss in UDP.

Minor cases of Packet Loss may not be of high priority, but extensive packet loss can cause frustration for users, and can exponentially grow to a larger problem if not monitored. Data centers and ISP providers are aware that packet loss has no single cause and can occur in every network, so they set a standard for the allowable percentage of packet loss. According to Cisco recommendations, packet loss on VoIP traffic should be kept below 1%, and between 0.05% and 5% depending on the type of video.

Troubleshooting Packet Loss

Packet loss troubleshooting cannot be done without identifying first that it occurs. Potential signs of packet loss can include:

  • slow connections or failure to load webpages and services
  • High CPU utilization
  • over-utilized devices
  • faulty hardware/software
  • cyberattack

From a network administrators’ point of view, most of these issues can be identified via monitoring tools. To have a look at tools that can be used to monitor network utilization and traffic check out articles discussing that topic:

(Link to articles in datacentertools)

After diagnosing that packet loss occurs, it is also important to further verify it through a method using ICMP (Internet Control Message Protocol). This protocol includes diagnostic tools called Ping and Traceroute that can be used to check for Packet Loss. Ping works by consistently sending Ping packets (of various sizes), to a specified destination. Depending on the success of the Ping it can help determine if there is loss on the network. Once this has been identified, Traceroute can also be used to determine which hop in the path from sender to receiver is causing the packet loss.

Ping and Traceroute can be done using Command Line Utilities, however, there are tools with user-friendly interfaces that incorporate Ping and Traceroute. One example of these tools is SolarWinds Network Performance Monitor. It has the capability to isolate segment failures which are significant in identifying the part of the network where packet loss occurs.

As packet loss is a major problem for UDP transmissions such as VoIP, SolarWinds VoIP and Network Quality Manager is a useful tool due to its real-time WAN and VoIP QoS monitoring. This tool includes a visualization module that shows the paths followed by VoIP, along with the health of each node in color-coded statuses and focuses on network conditions important to successful VoIP delivery.

However, these tools can only identify ongoing packet loss or route failure. They cannot provide an analysis of the transferred data nor fix the reasons for packet loss.

Since there is no universal solution to packet loss, as the causes of it are varied, fixing it can be trial and error. Below is a list of the basic steps that can be used to fix packet loss.

  • Verify physical network connectivity – cables and ports must be properly connected and installed, check that cables are in good working condition
  • Hard Reset – Restarting network hardware such as routers or switches can stop technical faults however, this is a service affecting solutions and needs proper implementation.
  • Using cabled connections – using cable connections such as Fiber optics rather than wireless can significantly improve connection quality
  • Software update- Keeping the operating system of network devices up to date will help ensure that no bugs in the OS are causing packet loss
  • Hardware Upgrade – Upgrading network infrastructure to get rid of inefficient and faulty hardware that cause packet loss
  • Quality of Service (QoS) setting implementation – Services that are traversing the network should be prioritized according to the most important
  • Increase Bandwidth- congested links can be addressed by increasing the allowable bandwidth of the link or upgrading it to a high capacity pipe to accommodate more traffic
  • Firewalls and IPS(Intrusion Prevention System) – cyber-attack can also cause packet loss via buffer over flow or DDoS, so installing Firewalls together with IPS can prevent or mitigate it
  • Verify Configuration – check configuration on network devices

 

Being able to swiftly diagnose and identify the reasons that cause packet loss is the most effective way to avoid it within a network. Although packet loss is not always critical all network devices become out of date over time and hardware components go through wear and tear. This should be fixed quickly and with the appropriate approach before it becomes a more critical network problem. Keep constant tabs on your network equipment using proper tools to ensure a healthy and effective network for your users.