A common method to efficiently monitor networks is by using Network Taps. These completely passive devices are inserted directly into the path between the network devices (switches, routers, etc.) and subtract a part of the signal in order to send a copy of this traffic to the right tools for further analysis. A tap (test access point) provides better accuracy especially under moderate to heavy traffic conditions without the limitations that often occur when using a mirror or SPAN port.
A TAP is essentially an optical splitter which divides the signal coming from both the send and receive data streams simultaneously on separate dedicated channels to ensure all data arrives at the monitoring device or packet broker in real time.
Figure 1: How a TAP sends and receives data
In figure 1 above, Network Device A sends and receives data from Network Device B. The TAP which is placed in between these two devices is receiving 30% of the total light to send to the tool side for each direction.
What is a Split Ratio?
A split ratio is the amount of light that is re-directed from the network to the monitor ports. The amount is expressed as the ratio between the percentage of energy which goes to the network, and the percentage amount which goes to the tool. For example, if a split ratio is 50/50, this means that 50% goes to the network and 50% goes to tool. Common split ratios are 50/50, 90/10, 80/20, 70/30 and 60/40.
Occasionally we’ll come across situations that have taps with multiple outputs per ingress port, where there’s a multiple split which leads to a distribution of energy in more segments. For example, 80/10/10 means 80% goes to the network, 10% to one tool and 10% to the second tool.
It’s important to note that the first percentage is always higher because you will want the network side to have the most power versus the monitoring side in order to preserve the network path as much as possible. This value can be a good indicator of approximately how much power will be transmitted to the monitoring ports and to the network with respect to the incoming power. Essentially, the power at the output would be diminished by these percentages. However, one should note that the fiber connections and other dispersion occurring inside the tap could cause additional power loss which must be considered when performing this calculation.
Figure 2 below shows a table which indicates the power loss (insertion loss) caused by the tap along with additional factors such as connectors.
Figure 2: Table showing optical fiber insertion loss
Sensitivity of the Receiver
When choosing the right tap and split ratio, it’s important to consider the transmitted power and the minimum power required at the receiving side. The transmitters at both ends are sending light with a specific optical power, which varies from transceiver to transceiver. The receiving side can reconstruct the information transmitted if it receives sufficient power.
The sensitivity of the receiver will indicate how weak an input signal can get before the bit-error ratio (BER) exceeds some specified number. In other words, it’s an indicator of how good a receiver is when there are low levels of incident optical power. In simpler terms, the more sensible the receiver, the more attenuation it can bear.
Calculating the Loss Power Budget
The link budget is accounting for all the gains and losses from the output of the transmitter to the input of the receiver. These gains and losses include: gains from all optical amplifiers, propagation loss due to the physical media, insertion loss due to the connectors and loss through any optical attenuator and/or optical splitters. Therefore, you should consider the impact on the calculation of parameters such as: link distance, fiber type, launch power and number of interconnections.
Optical power is calculated in decibels (dB), where X(dB) = 10Log X, and the calculation of this logarithm can help us simplify the calculation. In fact, any ratio between values becomes a difference in dB as well as multiplying becomes a sum.
The budget calculation is fairly straightforward as each of the components in this calculation is affecting the original power by dividing it from its initial 100% and subtracting from the initial total transmitted power. If you know Connector Loss, Propagation Loss (due to the fiber and proportional to the cable length) and Insertion Loss (which is the sum of the loss due to the splitter, dispersion and the connectors inside the TAP), all you need to do is subtract each contribution from the initial transmitted power. The resulting value must be greater than the sensitivity of the receiver.
Figure 3 shows a path of light through cabling and tap indicating various losses
P0 = initial transmitted power
L1, L2, L3, L4 = power loss due to the fiber which is expressed in dB/Km (characteristic of the fiber) and multiplied by the length of the fiber
C1 = connector loss
I1, I2 = insertion loss
Then the two endpoints will receive respectively:
Which split ratio is right for my network?
As mentioned above, to determine the correct split ratio, a loss (power) budget should be calculated. Choosing the right split ratio may seem complicated, but it’s actually relatively easy.
For starters, it’s important to differentiate between Single Mode (SM) optical receivers and Multi-Mode (MM) receivers (you can learn more about this difference in our blog here). SM optical receivers have higher sensitivity, so they can tolerate greater attenuation. However, pay close attention to the cable length between the two endpoints because if the distance is too long, the monitoring side can suffer lower power as a result of higher attenuation due to the longer distance.
The Multi-Mode (MM) transceivers are less sensitive and therefore lower split ratios are recommended to avoid too many errors on the monitoring side as a result of lower power.
For starters, we recommend an evaluation of a 70/30 split ratio for the SM and 50/50 or 60/40 for the multi-mode. Towards the end of a 50/50, which is the worst split ratio, this means that half the signal is sent to the network path. In dB, this is equal to approximately 3dB (-10*Log0,5 of pure splitting attenuation) on both sides, compared to a 70/30 split ratio which is around -10*Log (0,7) =1,5dB of attenuation on the network side and -10*Log (0,3) =5dB on the monitoring side. Clearly, there’s a larger advantage on the network side with a 70/30 versus a 50/50 which is 3-1,5 =1,5 db, respectively.
There’s a common misconception that 80/20 is a better choice to preserve the network, but this isn’t always true since -10Log(0,8) = 0,9 dB of attenuation will be on the network side but a good -10Log(0,2)=7dB will be on the monitoring side. Furthermore, considering that -12dBm is a normal sensitivity for standard transceivers, it’s clear that one must be cautious before making a final choice.
Ultimately, your choices depend on how sufficient the power budget is in order to satisfy the power requirements of both ends (network and monitoring) with any split ratio. Once this has been determined, we have several options to choose from: attempt to minimize the number of connections and/or length of the cables or choose more performing transceivers that have higher sensitivity.
Whatever choice you make, we recommend understanding your network needs, the location of where the taps will be, desired split ratio and light budget requirements.