What is a data packet?
A data packet is a unit of data that can be transmitted via a network, like the Internet. In fact, the entire Internet Protocol (IP) is based on data packet transmission. To send an item like a photo, for example, it is broken down into smaller pieces (packets) and sent over a network connection where it is assembled at the receiving end of the transmission, which will then display the photo. But sometimes, the photo doesn’t is missing pixels or doesn’t display at all. This can be the result of packet loss.
What is packet loss?
Occasionally, data packets traveling through a network get bumped off the transmission before reaching the final destination, this is known as packet loss. There are several reasons why this inconvenience could occur, including:
One of the most common causes of lost data packets, link congestion occurs when there are more data packets traveling a network link than the link is designed to handle.
When a device receives packets at a rate that exceeds what it can process (or the device’s buffer is full), those packets may simply be dumped off the device.
The type of network can have a significant impact on the quality and consistency of packet transmission; generally speaking, wireless networks are impacted more than wired networks. For example, radio frequency interference, weak signal, distance limitations, and roaming can all produce increased packet loss in wireless networks. Alternatively, in wired networks the main culprit of packet loss is usually faulty cables.
An enterprise network that serves thousands of employees may contain hundreds of pieces of equipment; if one device is configured to mismatch with another device, packet loss could occur between those two components.
Denial of Service attacks can result in legitimate data packets being pushed off a network if the network is being overwhelmed with attack traffic.
The Effects of Packet Loss
The specific effects of packet loss may vary depending on the protocol involved. TCP is generally designed to manage this issue; in that it acknowledges the receipt of packets and will attempt to re-transmit if that acknowledgement isn’t received. On the other hand, UDP lacks re-transmission capabilities and is poorly-suited for handling lost packets.
Similarly, the effect of packet loss also varies depending on the application in use, with video conferencing apps (like Voice over IP) suffering the most. Applications like Skype, Microsoft Teams, and WhatsApp are notoriously poor at handling dropped data packets and the video degradation will manifest as robotic speech or completely dropped audio.
Packet loss on VoIP traffic should be kept below 1% and between 0.05% and 5% depending on the type of video.
But regardless of the protocol or application, the ultimate effect of packet loss is a poor user experience.
How to Fix Packet Loss
Resolving packet issues within a network primarily involves identifying the cause (diagnosis) and then fixing it. Here are three different tactics for fixing packet loss.
1) Get a Fatter Pipe
If a network link is congested, widening the link (pipe) will allow you to push more traffic through it. This can be accomplished by adding hardware to boost throughput or increasing network buffers with software.
2) Prioritize Critical Traffic
Leveraging tools that provide Quality of Service (QoS) features will allow you to set priorities for mission critical traffic and avoid a poor end-user experience. For example, many enterprises that rely on video conferencing utilize NetMotion Mobility to prioritize that traffic to provide seamless high-quality VoIP.
3) Upgrade Devices
At the end of the day, devices that are over-utilized beyond their capacity may simply need to be upgraded to sufficiently process data packets at the speed required for optimal performance.