Ethernet -over-Copper to Fiber converters (media converters) are rather simple devices with no complex additional settings. Meanwhile, they have their own application specifics that must be considered when building a network.
This article includes recommendations on application and choosing media converters, operation principles of “Link Fault Pass-through” and “Far End Fault” functions, as well as converters behaviour in redundant networks.
LFP and FEF
One might think, that the sole purpose of a media converter is to convert a signal fr om one transfer medium (copper wire) to the other (optical fiber). This is true, however, modern media converters also support the functions that, if applied correctly, can enhance network reliability.
Needless to say, that the basic role of a media converter lies in data transfer between two devices that can’t be connected to each other directly. The converter must remain invisible to network devices. Basically, a media converter must “simulate” a cable. That may sound rather simple, however, when used together, such “cable” actually comprises two copper cables and two fiber optic cables (see Fig. 1 below). For that reason, there appeared such functions as “Link Fault Pass-through” and “Far End Fault”.
Link Fault Pass-through (LFP) function serves to notify a neighboring converter on a copper cable break. Let us consider, what will happen if a copper cable that connects the converter on the left breaks accidentally (see Fig.2). In this case, the converter on the right will not find out about the breakdown and, despite the connection disruption, the network will continue sending messages to “nowhere” assuming the connection is still active. Here is wh ere the Link Fault Pass-through (LFP) function comes in. The detailed operation of LFP is given in fig.2 below.
- Figure 1: Normal operation of two media converters.
- Figure 2: Copper cable up to converter A is broken. Converter A sends a special LFP message to converter B.
- Figure 3: Converter A switches off connection with converter B.
- Figure 4: Converter B switches off copper connection.
- Figure 5: Converter B switches off connection with converter A.
Far End Fault (FEF 802.3u) – function serves to notify on a breakdown of one of the optic fiber cable cores (!).
Let us consider, what will happen, if one of the optic fiber cable cores connecting the media converter on the left with the one on the right breaks.
In this case, the left converter will not be able to transfer data to the right converter. However, if another optic fiber cable core is still safe, the switch on the right will continue transferring data to the switch on the left, which can lead to transmission errors within the whole network. Here is where the Far End Fault function comes in. The detailed operation of FEF is given in fig. 3 below.
- Figure 1: Normal operation of two media converters.
- Figure 2: One of optic fiber cable cores going from converter A to converter B gets broken.
- Figure 3: Converter B switches off copper communication.
- Figure 4: Converter B switches off the second optic fiber cable core going to converter A.
- Figure 5: Converter A switches off copper cable connection.
Please note: use media converters in pairs
The illustrations above show that media converters are used in pairs. Many design engineers ignore this condition and use only one Ethernet-to-Fiber converter. The problem of using only one converter lies in the fact that, if activated, “Far End Fault” or “Link Fault Pass Through” function will not work properly. Not only media converters must be applied in pairs, they must come from one manufacturer and belong to one model, preferably. This follows from the fact that various manufacturers may use proprietary protocols for “Far End Fault” and “Link Fault Pass Through” functions. Moreover, one manufacturer may use different microchips in various models that will lead to interoperability between various models.
Let us consider what will happen if only one converter with LFP function is used (see fig. 4 below). Actuality, it is easy to realize why application of LFP function causes problems in this situation. Firstly, one should keep in mind that LFP is a function that is implemented within a media converter. If the connection over a copper cable between the left switch and the media converter gets disrupted, the media converter sends a message to the right switch saying that the transfer has been unsuccessful. The problem is that the switch will not understand an LFP message and, so, it will be dismissed. Due to this reason, the switch will continue transferring data to the media converter assuming data to be transmitted to the switch on the left. The fact that this is not happening can lead to network failure.
Operation of media converters in redundant networks
How the media converter operation mode affects the work of a redundant network
Redundancy is an essential aspect of any network and is crucial for industrial networks. Many Ethernet switches support STP/RSTP protocols, as well as proprietary protocols, for instance, Moxa Turbo Ring.
As we know, in order to secure the operation of redundancy protocols, the switches exchange with special service messages – BPDU.
Some of media converter models operating in Store-and-Forward mode are not able to transfer BPDU packets of redundancy protocols. This problem can be solved by switching converters into "Pass-through” mode, as in this mode optical fiber line ensures transparent communication line.
How the media converter affects network recovery time.
When using media converters in Ethernet network that applies STP/RSTP or Turbo Ring redundancy protocols, you should consider the way how the converters affect recovery time.
Let us consider each case individually:
- Recovery time for STP/RSTP is rather low, and, even if switching on of media converters will slow down operation a little, they will not affect the recovery time significantly.
- Proprietary protocols are much faster than STP/RSTP. For example, Turbo Ring by Moxa features the recovery time of under 20 ms. So, if you use media converters in such a network, you should better await significant recovery time reduction.