Packet loss is the amount of data that never reaches its destination. Typically, this means there is some type of problem with your connection which could be anything from a loose cable, faulty switch port, overloaded router slowing down traffic, a power outage at the modem, etc.
Many things can cause packet loss, so it’s essential to identify what is happening before committing resources to fix it. One easy way to find out how much packet loss you are really experiencing is by going to this link, a web-based tool to check for packet loss created by DeviceTest. They also provide other helpful web-based tools to check for the quality of your internet connection.
What is Packet Loss?
When you send out data through the Internet, it travels across your local network to your ISP, which is forwarded to the destination site of your choice. Packets consist of a header and a payload. The former includes information such as the source, destination, size, and time stamp. On the other hand, the latter contains whatever type of data you’re sending (i.e., an email or a webpage).
For example, when you send out an email from Gmail to your friend, it’s broken down into packets before being sent across the network to Google’s servers. Then it’s reassembled based on the information in the header before reaching its final destination.
In this regard, packet loss occurs when one or more packets fail to reach their destination because they get lost along their journey due to your fault. And while small amounts of packet loss may not seem like a big deal, it can lead to significant issues with your Internet connection.
Signs and Symptoms of Packet Loss: Connection Issues
It’s possible for packets containing data from other sources on your network to interfere with those that you’re sending. While this doesn’t impact the contents of those packets, it does affect the speed at which they reach their destination.
You’ll find this especially true if someone else in your household is downloading large amounts of data as well as when you try streaming videos or playing online games.
This issue increases as more packets fail but also depends on how quickly your ISP responds in delivering missing ones (i.e., upload speed). For instance, if one out of every 100 packets fails to reach its destination, you might experience problems loading a webpage. However, if 10 out of every 100 fail to arrive at their destination, you’ll likely lose the entire webpage instead of only part of it.
It’s also possible for packets from your ISP to interfere with those coming from the website itself. The former contains information about your connection, while the latter displays whatever type of content you’re trying to access.
In this regard, packet loss occurs when one or more packets from either side interfere with each other. This is especially true when other users are using bandwidth on your network and during peak hours when everyone is online.
Common Forms of Packet Loss
Packet loss is one of those issues that can cause this type of behavior. There are many forms of packet loss, but the two that will be covered here are frame loss and congestion control algorithms that try to compensate for them.
Frame Loss
Frame Loss is when some packets are dropped by the underlying sub-networks physical layer (LAN, WiFi, etc.). The amount of data sent over these layers is usually much more significant than the link can support, so some level of prioritization or error checking has to be implemented to ensure higher priority information gets through.
Congestion Control Algorithms
Congestion Control Algorithms are used in an attempt to combat Frame Loss on modern networks. A dropped packet in a TCP session results in retransmission by the sender. This is costly to both sides of the connection. The sender wastes time waiting for a timeout before resending, and the receiver has to wait for another entire round trip before it can get an update on where things are at.
The congestion control algorithms try to address this problem by monitoring how many packets get dropped and adjusting accordingly. Instead of waiting for a timeout after each dropped packet, they will increase their transmission rate over time if they detect that packets are declining rapidly. If they don’t drop or drop slowly then they will continue transmitting at the same rate as before.
Many congestion control algorithms try to address this. Some of the most commonly used ones are listed below:
New Reno: This is one of the oldest algorithms still in use today. It’s a bit more aggressive than others, so you will have to experiment with it if you choose to implement it.
Westwood+: This algorithm is part of what goes into making something like VOIP work well on your network. It tries to avoid filling up queues at either end while still transmitting as much data as possible while using less energy for transmission.
BIC: This is short for Binary Increase Congestion control and was created by the same people who wrote CUBIC. It’s a good choice for avoiding congestion due to large downloads or streaming video.
CUBIC: Another very popular choice is CUBIC which has been around since 2007. It uses Westwood+ under the covers and considers whether the link is symmetric in nature or not. If you are running over WiFi, this can be a big issue, so it might be worth looking into if you find that your network speed often fluctuates when watching online videos.
Aim DC: If the streaming video is your main concern, this might be your choice. It will throttle back the packet loss on high bandwidth links while not impacting low bandwidth links that do not need to transmit as much data.
H-TCP: This modification of TCP avoids some of its previous issues, like reordering packets during congestion. It accomplishes this by changing the way the sender chooses its transmission rate, which helps prevent some types of problems with dropped packets.
