Packet loss is one of the key factors that can negatively affect subscriber quality of experience (QoE), along with bandwidth and latency. It occurs when one or more packets don’t reach their destination, and can seriously impact network performance and data quality. In this article, we’ll take a look at what causes packet loss, its effects on network QoE, and how to detect and reduce it where possible.
What is Packet Loss?
First, let’s start by defining our terms. A packet is a small unit of data that’s transmitted between origin and destination points on a network via a transport protocol. All data sent over the internet (e.g. videos, emails, images, etc.) is comprised of packets. These are broken down into smaller packets first to make the data easier to transmit, then reassembled once they reach their destination.
Packet loss refers to when one or more of those smaller data packets don’t get to their destination. This leads to degraded network performance, increased retransmissions, and reduced data quality in real-time applications, such as online gaming or VoIP calls. Slower service caused by packet loss negatively affects both residential and commercial customers, and is likely to lead to an increase in tickets and calls for your support team.
What Causes Packet Loss?
Packet loss has a variety of causes, including faulty devices or physical transmission issues, such as interference or damaged cables. One of the most common causes, however, is congestion caused by network bottlenecks.
Examples of bottlenecks include overloaded access points, low signal strength CPEs, or in-home WiFi issues. A bottleneck is essentially a piece of equipment that has more data coming in than it can transmit, causing it to start buffering packets and eventually dropping them. This causes an increase in latency, and eventually packet loss, leading to a slow-feeling internet for end users.
The Impact on Networks
It’s useful to also consider throughput when thinking about the impact of packet loss on a network. Both packet loss and throughput are crucial for assessing network performance and reliability. By understanding and monitoring these metrics, core network operations teams can optimize data transmission and enhance the user experience.
Throughput measures the amount of data that can be transmitted over a network within a given time frame. It reflects the network’s capacity and efficiency in data transfer. This is influenced by available network bandwidth, representing the maximum data rate a network can handle. However, achieving maximum throughput is often limited by factors like network congestion, latency, and packet loss.
Higher packet loss typically results in lower throughput as retransmissions reduce the effective data transfer rate. As well, excessive packet loss negatively affects response times, audio/video streams, and overall application performance.
A smaller increase in packet loss has a significant impact on throughput. Looking at the graph, we can see that once packet loss reaches 1.0%, the throughput drops dramatically. This behavior is commonly observed in TCP, where higher packet loss leads to reduced throughput.
How to Reduce Packet Loss
As the saying goes, an ounce of prevention is worth a pound of cure. In other words, the best way to reduce packet loss and its negative effects is to use systems that proactively prevent it in the first place, as opposed to trying to mitigate things after the fact.
As mentioned above, packet loss can be caused by network congestion. An example of this would be that when a link gets saturated (i.e. at more than 80% capacity), all the buffers associated with that link also become full. This creates bufferbloat, which in turn causes an uptick in packet loss (and latency), leading to poor quality of experience for your subscribers.
Preseem includes scores and other analytic features that are designed to help operators proactively address capacity constraints and other network problems that cause packet loss. We also employ active queue management (AQM) techniques to maintain a good experience for subscribers, with low latency and loss, even when networks are at their busiest.
Of course, packet loss can’t be completely eliminated and is a normal part of any network. Acceptable levels of packet loss vary according to the data being transmitted. For example, Cisco recommends acceptable packet loss guidelines of 0.5% for video and up to 1.0% for audio.
As we’ve seen, however, packet loss and its negative effects can be reduced by adopting proactive QoE systems like Preseem that empower you to minimize network congestion before it starts affecting your customers.
For more on what causes packet loss, as well as discussions on low latency, bufferbloat, and more, listen to the Preseem Podcast or check out check out this episode recap of when our team chatted about these topics with networking expert Bjorn Ivar Teigen, Head of Research at Domos.