A network administrator has the job of monitoring the performance of his or her network. That means the Internet connection, the Wi-Fi network, ping, and more. As part of that monitoring responsibility, an administrator needs to know how to stop packet loss. However, to first do that, a network admin must know what packet loss is.
What is Packet Loss?
When a computer device accesses the Internet or any network, it uses packets. These small packets of data get sent between the network and the device. If one of these packets fail to get to its intended destination, then it suffers packet loss. Packet loss can take many forms; some examples include:
● Network disruption – The Internet cuts out on occasion, somewhat randomly, before returning to a stable state
● Slow Internet service, such as videos constantly buffering or webpages loading slowly
● Total loss of network connectivity, meaning devices cannot connect to the network, not even briefly.
The Different Types of Packet Loss
An Internet packet loss, also sometimes called latency or lag, may appear at different points along the connection. Latency refers to the time it takes for data to get transferred between the sender and receiver. In an ideal world, that time should feel instant for a user, which would be considered a low latency connection. However, the loss of those packets via a high latency will ruin a network’s performance and thus a user’s online session.
On a more local level, Wi-Fi packet loss can occur. The packets used by the wireless network get transferred over the Wi-Fi signal, with any disruption causing data packets to get lost. Wi-Fi packet loss usually gets caused by overloading the network or the signal delivering the data packets getting disrupted, hence only delivering partial packets. Some more examples of what may cause packet loss, are detailed below.
What Causes Packet Loss?
Packet loss has no prejudice when effecting websites and applications. However, it often appears in applications that use real-time packet processing, or UDP processing, with a further explanation on that further below. Video, audio, and gaming applications are all susceptible to packet loss because they need the data right away, without much consideration of how it will get to the devices. Many businesses also use real-time packet processing, and in a business environment, even a brief period of packet loss can ruin a working day, as it can mean the company’s website or application isn’t working as intended. But how does packet loss happen?
Network congestion is the most common reason for packet loss. All networks, whether Wi-Fi or LAN, whether using the best network infrastructure in the business or not, are all susceptible to overuse. If a network is running over its maximum capacity, so has more users than it can handle, any data packets getting sent will have to wait in a queue. As the network attempts to catch up with the packets in the queue, some will get lost, answering, “What is packet loss?” at least in one example. More modern, robust systems do have the ability to circle back and collect these dropped packets, though that is not always the case.
Issues with Network Hardware
Faulty, old, and generally glitchy hardware can weaken a network’s capability. As stated, a top of the line network can still suffer network congestion, though a system that is already having issues, is expected to undergo some packet loss. A business, especially, should always take measures to keep its network up-to-date or expect a lapse in network performance.
Bugs in the Software
Sometimes the glitches found in the network hardware, as mentioned above, can get fixed. These bugs in the software of the network can disrupt the performance of the network. This disruption ultimately ends up leading to the loss of packets. On occasion, a simple reboot of the network can eliminate these software bugs. However, if a bug has found its way into the system due to a hardware update, then a network administrator should avoid updating to that version until the product provider patches it. Many software providers provide a list of known bugs after the patch gets released, leaving the decision to the network admin as to whether to proceed or wait for the next one.
Network congestion has already been explained and should not be confused with the possibility of overloaded devices. A network can only handle so much traffic and a device can only manage so many processes. If a device gets overloaded with incoming packets, with packet size being a contributing factor, any new packets will struggle to get through. Again, like a network, any packets attempting to get processed will join a queue. However, also like a network, the queue can get too full, which leads to packets in the queue then getting lost due to the device’s inability to manage the incoming data.
Third-party Security Threats
On occasion, a packet loss is not down to an error on the network side; instead, coming from an external third-party. One of the more popular third-party attacks comes in the form of packet drop-attacks. A packet drop-attack entails a hacker getting into the router of the targeted network and then telling it to drop multiple packets of data. This sudden surge of data will substantially reduce a network’s ability to process packets, thus slowing down the system.
Another form of third-party attack comes in the form of a denial-of-service attack, or DOS attack, for short. A DOS attack aims to stop legitimate users from using the website or application. A hacker floods the network with traffic, to the point that it can’t handle anymore, leading it to crash. Any packets that aimed to get to the network previously will be lost, hence leading to packet loss. If a system is under attack, there is not too much a network administrator can do. However, if they can work quickly enough, they can block the offending IP addresses, halting the attack, though not always indefinitely.
Sometimes, it is not the network, nor a third-party attack that leads to packet loss. Sometimes, the infrastructure in place, to avoid network packet loss, is not sufficient to handle the job. Many network administrators build their monitoring system for the network, using a variety of third-party tools to meet their preferences. These tools get pieced together to do a job they weren’t initially created for, leaving a network ill-equipped to handle packet loss. With that being the case, a network administrator loses the tools that a pre-built dedicated system would bring, and thus the ability to manage packet loss efficiently.
The Relation Between Ping and Packet Loss
As addressed above, packet loss happens when the data packet can’t get to its intended destination. When this happens, as also discussed above, the Internet connection isn’t sufficient enough to get the data across, along with many other reasons for packet loss. However, if the packet loss isn’t happening regularly, but occurs at random intervals, there are three things a user can check, to help explain the packet loss. Many third-party websites can offer this information with a simple search of “Internet speed.”
The upload speed is the measurement used for when a user uploads data. Uploading of data includes sending files via email or using a video or audio chat system. Upload speed typically gets measured in MBPS or megabits per second.
Download speed, as the name implies, is the opposite of the upload speed. Download speed refers to how fast data packets are downloaded from a server, whether that is from visiting a website or from downloading media, such as a video or audio file. By design, download speed from an ISP is faster than the upload speed, though, as, with the alternative upload speed, the download speed still gets measured in MBPS.
Ping refers to the Internet connection’s ultimate speed when a user sends a request to a network. The faster the ping packets get delivered, the more responsive the connection, which is vital for applications such as voice and audio chat apps and online gaming. Ping gets measured in ms or milliseconds. Ideally, anything below 20ms will provide a near-instantaneous reaction, whereas anything over 150ms will come with noticeable lag that will ruin most Internet sessions. However, no matter how fast the ping is, there is still a chance for a packet loss issue to occur. Just because the connection is established and forwarding the data quickly, the avenue the data takes can cause packet loss.
How Can You Reduce Packet Loss?
The above details the most common reasons for a network suffering packet loss, and though the solution to some of the examples is apparent, the others are not. Unfortunately, no matter what measures get taken, there is no way to eliminate packet loss completely. As stated, no matter the system, it will always be open to becoming overcrowded, and sometimes, a third-party can strike to ruin the day. However, just because there is no way to eliminate packet loss, it doesn’t mean that the chances of a high packet loss level can’t get reduced.
Check the Network Connections
It may seem like a simple step, but if the connections are not stable, then a business should expect some form of packet loss. If using a cabled set-up, all cables should be securely connected and the wires should be damage-free. If relying on a Wi-Fi connection, the signal should be strong. A user working with a single bar of signal strength is asking for packet loss.
Restart the System
Many people use a system, without ever turning it off. However, this can be detrimental to a network. It is always a good idea to reset routers and other networking hardware, which allows them to update to the latest patch version. These patches help to remove any small bugs or glitches.
Swap Out Wi-Fi for Wired Connections
As stated, if the network is wireless, a stable Wi-Fi connection is necessary to avoid packet loss. However, even then, there is no guarantee that packet loss can get avoided, as there is no physical connection to carry the packet, with the packet instead traveling through the air via the wireless connection. In an ideal world, a fiber-optic connection is best, though even a standard ethernet connection is better than relying on a wireless connection.
Remove Devices That Cause Static
If a network is suffering packet loss, it may be due to peripheral devices, such as network cameras, Wi-Fi speakers and headphones, and anything else that connects to the network, but isn’t necessary. Audio devices, especially, cause RFI or radio frequency interference. Radio signals can interrupt Wi-Fi signals, causing a loss of packets. Multiple firewalls can also cause issues as they require a lot of bandwidth. A system only needs one firewall and that one can be disabled for a brief period if there are signs of packet loss.
Keep Software Updated
Many users often don’t restart their system, but another sin is hitting “remind me later” whenever there is a prompt to update the software. An operating system and web browser have updates for a reason, with that reason usually being to fix bugs and improve the user interface. Software that is regularly updated is much more likely to avoid packet loss.
Keep Hardware Updated
The above also applies to the physical hardware of a system. The equipment of a network should remain updated to ensure it is working at maximum efficiency. Faulty hardware will ruin a connection, and even then, if older devices are working as intended, upgraded network devices appear regularly. A business would do well to make use of these new products, and thus keep its system updated.
Make Use of QoS
QoS or Quality of Service settings allows a network admin to avoid packet loss by prioritizing specific devices on a network. A QoS is often necessary when dealing with a system that transmits resource-intensive data, such as streamed content, online gaming, VOIP calls, and video calls. A QoS prioritizes such systems to avoid packet loss.
Are You Suffering Packet Loss?
The above makes it clear what causes packet loss, though, to an amateur network administrator, it can be challenging to understand if packet loss is happening or not. The results of packet loss depend on the application or protocol and the effects can differ. A TCP protocol, for example, or transmission control protocol, accepts data packets. It then attempts to retransmit any data that gets lost on the way, meaning not many will realize packet loss has taken place. On the other hand, a UDP, or user datagram protocol, will not retransmit data, leaving it lost forever.
The best way for a user to understand they are suffering packet loss is when it happens during a VOIP call session via services such as Skype, Facebook Messenger, Discord, and similar. If packet loss occurs, the call will effectively break up, as it would with a traditional phone call that can’t receive adequate cell service. The quality of the call dropping or signs of stuttering in the conversation are a couple of indications that the network is suffering packet loss. If the data making the call possibly gets lost because it is relayed real-time, it can be difficult for the application to recover. As little as 5% of discarded packets can ruin a VOIP call.
What is a TCP Protocol?
As stated above, a TCP protocol will accept data packets, and if there is a case of packet loss, it will try to forward the packets again. However, to understand how that works, it is essential to understand what a TCP protocol is. TCP or TCP/IP is a collection of rules or protocols that allow computers to send data over the network. It is the most common form of transferring data, with the protocol initially coming from the U.S. Department of Defense. The DOD wanted a way to transfer data that prioritized accurate delivery when data gets transferred between two points. It is for this reason why data packets will get re-sent if they initially end in packet loss. To further ensure accuracy, the newly sent packets will attempt to take a different route, if the ones sent before, get lost.
However, despite the accuracy of a TCP connection, there are security concerns that come with the protocol. It is for this reason why it is ill-advised to use public Wi-Fi networks. A VPN or virtual private network can bring additional security to secure data transfer. Additionally, if packet loss gets caused by lackluster equipment at the ISP’s end, the servers of a VPN service can help rectify the poor connection.
What Is a UDP Protocol?
Using a TCP protocol is a great way to avoid packet loss. However, users don’t usually get to choose how their data gets handled, with the application or website sending the data being in control of the data transfer protocol. UDP is another form of protocol that can send data quickly, though without the security features found with a TCP protocol, hence the moniker Unreliable Datagram Protocol. A UDP does not need to receive a TCP handshake, meaning the data getting sent does not need to be accepted, hence the emphasis on the speed of the transmission rate. One such example of an application that uses UDP packets is the above VOIP call, which can handle some form of packet loss, while still functioning. A real-time static-filled conversation is often preferable to a clear discussion that comes with a delay, as would be the case with TCP, which tries to send all data packets, even if they were initially lost. Applications that use UDP are incredibly open to DOS attacks.
How to Test for Packet Loss
Understanding the symptoms can only do so much. Packet loss is the most obvious suspect for a lousy quality VOIP calls, or gaming session, or general browsing experience, though, it might not always be the offender. For that reason, at the first sign of packet loss, a network administrator should perform a test on their system. The method for testing packet loss differs depending on the operating system in use, though all are very similar from the below test on a Windows OS.
Testing for Packet Loss on Windows
The Microsoft Windows packet loss test comes in two forms. One for testing for packet loss, due to a weak signal over a Wi-Fi network, and the other for testing for packet loss, due to network congestion. No matter which test is getting performed, the initiation of each is the same.
- The user performing the analysis needs to open up the “run” application. It is possible to search for this in the search bar of the start menu.
- With the run application open, the user needs to type in “cmd” and then open the newly revealed “cmd” application.
Alternatively, the application will open if a user holds down the Window key while pressing R, and then typing “cmd” into the new window. With the “cmd” app open, testing can begin.
Testing for Wi-Fi Coverage
- Before performing the test, the user and their device should be away from the network’s source. There should also be physical barriers, such as concrete walls, in between the two. Doing this will ensure that the test results are clear as to whether or not the issue is with the Wi-Fi coverage. Packet loss could stem from something else, such as the wireless devices not able to utilize the network efficiently or if the ISPs Internet delivery is lacking, and testing the Wi-Fi coverage will help narrow it down.
- With that accomplished, the user then needs to find the IP address of the source of the network. To do this, the network administrator needs to type in “ipconfig” into the “cmd” app, which will show the IP address.
- With the IP address discovered, a ping then needs to get sent to the router. Once more in the “cmd” app, the user needs to type in “ping [target IP] -n 25” with the IP address replacing the target IP field. Running the command will perform the test.
The test will then send 25 packets of data to the head of the network. Once those packets are received, the command will calculate the time it takes for the data to perform a round trip, along with the percentage of packet loss rate. Ideally, the round trip should take less than 300ms and there should be less than 1% of packet loss.
Testing for Network Congestion
Testing for network congestion takes a lot less effort than a Wi-Fi coverage test. To perform a network congestion test, a user should sit with their device next to the router. The user should then access a reasonably stable Internet site, such as google.com. If the device accesses the website with no problems, and almost instantaneously, then all is working as intended. However, if the user sits next to the router and the webpage still takes an unnecessary amount of time to load, because the data packets aren’t getting through, then network congestion may be to blame. The packets may come through at one point or another, though this delay causes a jitter, as it is known.
One resolution is to contact the ISP, or Internet service provider, and request more bandwidth. Alternatively, a network admin can make use of WAN optimization, which is the process of giving priority to specific parts of the network so that they receive more bandwidth. Doing so will cause some devices to suffer network congestion, but not those that are more important.
Use a Network Monitoring Tool
The above two tests are suitable for deciding whether a network is suffering from packet loss or not. However, the best way to both detect and prevent package loss is to monitor the system via the aptly named network monitoring tool. The tool provides a comprehensive view of the network, which allows a network administrator to address packet loss before it happens. Some pieces of software provide alerts when a system reaches capacity, while others will enable a user to set priority for a device, similar to how QoS works. A network monitoring tool gives an overall view of a network’s performance, which is indispensable information.
The Most Common Cause of Packet Loss is Complacency
Putting a network in place and then letting do its job without any monitoring is the recipe for packet loss. Understanding
what is packet loss is not enough. The world of networking technology is always changing and it is a network administrator’s responsibility to make sure that the network they monitor has up to date software and hardware. Even with the best infrastructure in place, a network admin needs to continually monitor network traffic to ensure every user is getting the best Internet experience by keeping missing packets to a minimum. And, if everything is working as it should, but there is still an occurrence of random packet loss, the problem should get tracked down and fixed, even if that means changing ISP.
In truth, combatting packet loss is a full-time responsibility, and a second of complacency can lead to a network getting overwhelmed, which ultimately leads to the packet loss problem.