Given a scenario, troubleshoot common physical connectivity problems. This chapter covers in detail different physical connectivity (cabling) failure scenarios, including bad wiring and connectors, split cables, dB loss, cable placement, interference, and crosstalk. We cover network troubleshooting in great detail in our Cisco CCNA course.
Troubleshooting Faulty Cable Connectors
Network connectors are an important part of the entire flow of traffic, as they are the beginning and the end of every communication link. If you have a problem with the connectors (Layer 1), you will have a problem at every other upper layer and the applications will not function.
Connectors are susceptible to problems because they are often used and moved within a network. Some parts of the connectors might become faulty when using them, as opposed to other network components, like wires, that never move from their initial position. Considering the large number of connector types (as presented in Chapter 15), there are many different ways a connector can fail.
When analyzing a connector (e.g., an RJ45 Ethernet connector), you should carefully inspect the following aspects:
- Verify that it is correctly crimped, including the order of wires
- Verify that all wires are touching the metal connector blades and that the crimp is not partial
- Verify that no wires are missing
- Verify that the plastic cable jacket is fixed into the connector (i.e., wires are not hanging out of the connector)
- Verify that the RJ45 lock is functional on the connector (otherwise, the connector will pop out of the port)
If any of the problems mentioned above are present, you should consider changing the connector to avoid Layer 1 issues.
Troubleshooting Short and Open Circuits
Generic wire problems are usually called short circuits. However, cable issues involve both short circuits and open circuits, which are completely opposite problems. A short circuit involves two wires within a cable or connector touching each other. In a normal operation mode this should not happen, as each wire in a cable is isolated inside a shell to prevent interaction with the other wires as it transmits an independent signal. If two wires do touch, the signals associated with each of them will cross from one to the other, causing a lack of signal.
On the other hand, an open circuit involves a complete disconnection of a wire. This might be easier to identify than a short circuit, as disconnected cables are easier to verify. Just as with a short circuit, an open circuit generates a lack of connectivity on the specific wire/cable.
Note: Sometimes a short or open circuit will cause only an intermittent lack of connection on a small scale. However, it is not likely that you will find a situation in which one of these issues generates a degradation of the signal. Usually it is either a permanent signal loss or an intermittent signal loss. |
Troubleshooting short and open circuits may be difficult, as these problems are often not physically visible. Sometimes if you move the cable you will see a signal recovery and this may put you on the right track. However, the recommended way to investigate such issues is using cable testing equipment that works with copper or fiber cables, as shown in Figure 18.1 below. Such devices can provide advanced results, such as:
- Which wire has the problem
- What the problem is
- How far the problem is (location between endpoints)
Figure 18.1 – Cable Testing Equipment
Solving short and open circuit problems usually involves replacing the cable or connectors, as such problems are almost impossible to repair on a per-wire basis.
Troubleshooting Split Cables
Split cables are basically cabling mistakes (usually in UTP cabling) that translate to specific wires being twisted in the wrong pairs. One of the challenges of troubleshooting a split cable is that a regular wire mapping test would show no errors, as each wire is connected to the proper pin at each end. In this case, the problem is within the cable and is hidden by the cable jacket.
Split cables usually impact performance as the wire gets longer because near-end crosstalk occurs and it increases with distance. As traffic is sent further down the line and the signal is subject to interference, you don’t have the correct twisted-pairs that would normally cancel that interference and ensure proper signal quality.
As every twisted-pair in a UTP cable can be considered a self-contained entity, the signal and performance will suffer if you misconfigure them and twist the wires in the wrong pairs. Let’s consider an example by analyzing a straight-through cable connecting a workstation to a switch. In a normal situation, the pairs should be similar to those in Figure 18.2 below:
Figure 18.2 – Straight-Through Cabling
Following the TIA/EIA-568A standard, the green wires (1 and 2 on the left image and 3 and 6 on the right) should be twisted together and should connect the TX on the left side with the RX on the right side. In a similar fashion, the orange wires (3 and 6 for 568A and 1 and 2 on 568B) should be twisted together and connect the TX on the right side with the RX on the left side. The other wires are usually not used in a FastEthernet environment.
Now imagine that the green and orange wires are still connected to the same pins but they are not twisted with the other same color cable and instead are twisted with the other color, meaning the green/white wire is twisted with the solid orange wire and the solid green wire is twisted with the orange/white wire. This is a split cable effect that does not impact connectivity but does lead to performance degradation, as it does not offer cancelation between signals on the same pair.
Troubleshooting dB Loss
Signal loss is an issue in both copper and fiber cabling infrastructures. This is represented by the percentage of signal lost at the destination based on the initial transmission sent. In other words, the signal received is not as strong as the signal sent.
Signal loss is also called attenuation and it increases as the signal travels through the media, which can include:
- Copper (electrical signals)
- Fiber (light beams)
- Wireless (radio waves)
The amount of signal loss depends on a number of different criteria, including:
- Distance
- End connectors
- Cable quality
- Patches between the endpoints
- Electrical interference (for copper cables)
- Radio interference (for wireless traffic)
Note: Signal strength is measured in units called Bells (B) that are divided into decibels (dB), which are one-tenth of a Bell. The B is capitalized because the unit is named after Alexander Graham Bell, the inventor of the first practical telephone. |
Signal modification can be expressed in two ways:
- Signal gain: the signal received is higher than the signal transmitted
- Signal loss: the signal received is lower than the signal transmitted
Signal loss also involves using a logarithmic scale. This logarithmic behavior means that the signal degradation does not follow a linear scale based on the dB number; instead, it increases much faster. To exemplify this, let’s use the following examples: If you have a difference of 3dB lost between the original signal and the signal received, this means that the signal received is half that of the transmitted signal. If the difference is 10dB, this means that the signal has been degraded 20 times. At 20dB there is a 100 times difference in signal loss and at 30dB there is a 1000 times difference in signal loss.
In the case of optical fibers, when installing them over long distances, you should analyze the MMF or SMF cable specifications to find the rate of signal loss. For example, the signal loss specified by the cable manufacturer can be 3dB for each km. On top of that, you should add the signal loss on the patch panels, also specified by the manufacturer, which is usually below 1dB (e.g., 0.3dB). Each time you go through a patch panel you lose signal strength, so a fiber cable that travels across a 2 km distance and through two patch panels has an end-to-end signal loss of 6.6dB based on the figures mentioned earlier (2x3dB per km and 2×0.3dB per patch panel).
Depending on the initial signal strength and the receiving end requirements (e.g., applications, services, etc.), careful analysis should be done in order to calculate the exact signal loss over long distances. Based on the level of signal loss, a connection can have a number of symptoms:
- Lack of connectivity (as a result of severe signal loss)
- Intermittent connectivity (enough signal strength to bring up the link but not enough to maintain it)
- Slow performance
- Interface errors
- Data corruption
Dedicated link quality testing equipment can help troubleshoot signal loss issues and can provide useful information on the health of the physical media (i.e., copper, fiber, or wireless).
Troubleshooting Reversed TX/RX
One particular type of cabling issue is the one in which the Transmit and the Receive pairs of a cable are inversed so the TX sides are connected to each other and the RX sides are connected to each other (as opposed to the correct way of connecting TX to RX).
In an Ethernet environment, this can lead to a straight-through cable being transformed into a crossover cable, and vice versa. This scenario is easy to identify and might not be an issue if the devices can automatically detect and correct this. This functionality is called auto-sensing or auto-MDIX and is transparent to the user.
The mistake can be at the connector or somewhere in the patch panel, and this is usually a human error. Because the error is not somewhere inside the cable, it can be relatively easy to identify by visually analyzing the cable terminations.
These types of issues are also present in optical fiber installations and can come in two forms:
- The cable has dedicated TX and RX connectors that are inserted into the wrong ports on the device
- The cable has a single connector (with both TX and RX fiber strands) that is incorrectly crimped (TX at one side of the cable to RX on the other side of the cable)
Solutions to fixing reversed TX/RX issues include:
- Changing the connector
- Changing the patch panel termination
- Inversing the RX and TX connectors (i.e., optical fibers)
Just like in other types of cabling issues, to find the exact problem and location, you can use dedicated diagnostic equipment (e.g., wire mapping devices) that can provide accurate results on the cable and connector configurations.
Cable Placement
In addition to choosing the type of cable, another important factor you should consider is the way you install and place the cables in your infrastructure. You should concentrate on three different areas for this:
- Cables connecting workstations: can go through the ceiling, in the floor, or on the floor
- Cables between floors: in case you have to connect devices located on different floors within a building
- Cables in the data center/server room: focus on cable management, as all the connections in the building are concentrated in this place and if cables are not tagged and placed in order this might create chaos; the high density of devices in the racks imposes proper cable management to avoid Layer 1 issues within the entire network
Cable placement becomes very important, especially if you mix different copper and fiber cable types. Copper cables are usually a lot heavier than fiber ones so if you have a large stack of copper cables, this might crush more sensitive fiber cables positioned below them.
During the installation process, you should consider the cable category you are using and the way this will scale in the future. Whenever possible, you should use the higher category available to avoid cable replacement with future network upgrades. For example, even if you only need Cat 5 upon installation, a wise decision would be installing Cat 6 cables to support future needs. It is very difficult to remove old cables and install completely new cables in a large building.
To optimize cable placement, the data center wiring plant needs to be centralized. This is usually placed in the middle of the data center to maintain minimum cable lengths. With this placement, you can easily distribute cables wherever you need to.
Various providers offer structured cabling systems that are suitable for large environments. Such a system includes passive components that organize and label cables at multiple points in the infrastructure. When using a structured cabling approach, this allows you to spend as little time as possible troubleshooting Layer 1 issues so you can focus more on the upper OSI layers during troubleshooting. An example of a structured cabling system is presented in Figure 18.3 below:
Figure 18.3 – Structured Cabling System
Electromagnetic Interference
Once the cables are installed, you have to make sure they are protected from electromagnetic interference (EMI) or other types of interference that could cause problems with the signal that is traveling through the cable.
There are a few things you should consider when you first install a cable from a handling perspective to minimize negative interference effects:
- Do not overtwist the cable
- Do not pull or stretch the cable
- Pay attention to how and where you use cable ties
- Pay attention to the bend radius; different types of cables allow a different maximum bend radius, so read the cable specifications carefully
Whenever you are working with cables, EMI can cause issues, so when operating in such an environment, you might consider using fiber cables instead because they do not use electrical signals and thus EMI does not affect them. Here are a few things you should consider to avoid or minimize EMI when using copper cables:
- Avoid power cords
- Avoid electrical cables
- Avoid fluorescent lights
- Avoid fire prevention components
- Avoid any kind of powerful electrical sources
Once the cabling installation is complete and before taking the network into production, you should use dedicated testing equipment to make sure that you have a good connection on every link and EMI is not affecting your signal at any point.
Troubleshooting Crosstalk
One challenge when working with copper is the crosstalk effect. This happens when a signal passing through a wire causes interference to a signal on another wire. This might be a case of signal leaking.
A common method of measuring crosstalk involves using dedicated cabling testing equipment that sends a signal across the cable and then analyzes the crosstalk on the other twisted-pairs that form the connection. Crosstalk comes in multiple forms:
- Near-end crosstalk (NEXT): interference measured at the transmitting end. This is one of the most relevant measurements, as the signal is the strongest near the transmitting end.
- Far-end crosstalk (FEXT): interference measured at the opposite direction from the transmitter end (i.e., receiving end).
- Alien crosstalk (AXT): interference from other neighbor cables. This is not usually seen in a properly isolated network infrastructure but it should be considered an alarm when happening.
- Attenuation to crosstalk ratio (ACR): the difference between insertion loss and NEXT. The signal strength is measured at the insertion point and then measured based on NEXT to figure out how much of the initial signal has been affected.
In situations in which you have tested the cables and see a lot of crosstalk, you probably have a connection problem due to several possible reasons:
- Cable incorrectly crimped
- Low-quality cable (not enough twists)
- Issue at punch-down points
Note: To minimize crosstalk effects, Category 6 UTP cables have an increased cable diameter so the distance between pairs of wires is increased. This increases the isolation between the signals carried over each pair. |
Cable installation should be carefully tested before taking them into production to be sure that crosstalk or other issues do not affect end-to-end communication. Dedicated testing equipment is recommended during this process.
Summary
When analyzing a connector (e.g., an RJ45 Ethernet connector), you should carefully inspect the following aspects:
- Verify that it is correctly crimped, including the order of wires
- Verify that all wires are touching the metal connector blades and that the crimp is not partial
- Verify that no wires are missing
- Verify that the plastic cable jacket is fixed into the connector (i.e., wires are not hanging out of the connector)
- Verify that the RJ45 lock is functional on the connector (otherwise, the connector will pop out of the port)
Generic wire problems are usually called short circuits. However, cable issues involve both short circuits and open circuits, which are completely opposite problems. A short circuit involves two wires within a cable or connector touching each other. In a normal operation mode this should not happen, as each wire in a cable is isolated inside a shell to prevent interaction with the other wires as it transmits an independent signal. If two wires do touch, the signals associated with each of them will cross from one to the other, causing a lack of signal.
The recommended way to investigate cable issues is using cable testing equipment that can work with copper or fiber cables. Such devices can provide advanced results, such as:
- Which wire has the problem
- What the problem is
- How far the problem is (location between endpoints)
The amount of signal loss depends on a number of different criteria, including:
- Distance
- End connectors
- Cable quality
- Patches between the endpoints
- Electrical interference (for copper cables)
- Radio interference (for wireless traffic)
One particular type of cabling issue is the one in which the Transmit and the Receive pairs of a cable are inversed so the TX sides are connected to each other and the RX sides are connected to each other (as opposed to the correct way of connecting TX to RX).
Once the cables are installed, you have to make sure that they are protected from electromagnetic interference (EMI) or other types of interference that could cause problems with the signal that travels through the cable.
Cisco network troubleshooting guide.
Learn troubleshooting techniques and commands in our 101 Labs – CompTIA Network+ book.