The Class System is Dead, at least for optical fibre cabling.

for Netcomms Europe by Mike Gilmore, Standards Director

Those readers “au fait” with ISO/IEC 11801 and EN 50173 standards for generic cabling design will recognise the concept of cabling “channels” and “links” and the application of the term “Class” when referring to the requirements that apply to those channels.

Class has been applied to copper since the earliest editions of the standards and was extended to optical fibre cabling in 2001/2002. However, it has always been a bone of contention that the application of Classes to optical fibre was not directly equivalent to its use for copper – and was recognised as the best of a lot of worse solutions. Finally, it has been agreed to remove the optical fibre Classes from the next generation of the standards – unless someone comes up with a better idea which fixes the problems of the existing solution.

To explain the fundamental problem with the optical fibre Classes we have to review the multiple functions of the cabling Classes. Its first objective was for each Class to define a performance envelope within which one or more network applications could be grouped.

If we look at the copper Classes, any application of Class D would operate over a Class D channel. As each Class (A to FA) is hierarchical, a Class D would also support applications of Class A, B and C. As new applications were developed we had to create new channel Classes such as Class EA to support 10GBASE-T. On the odd occasion we drifted away from this concept by creating new Classes to reflect the then “state-of-the-art” cabling. Examples of this were Classes E, F and FA. This resulted in Classes without an application mapping – but this it not a fundamental flaw of the Class system. However, one critical feature of the copper Class concept was that the performance envelope was length independent – instead the performance envelope was defined by transmission parameters such as attenuation, NEXT and return loss. By comparison the optical fibre Classes were defined by attenuation and, indeed, by length. We started with OF-300, OF-500 and OF-2000 into which the existing applications were grouped and for which there was a hierarchy with OF-2000 applications being able to operate over OF-500 channels. So far so good!

However, the reader might be encouraged to ask why, if copper channels were not length dependent what made it necessary to use length for the optical fibre equivalent? The answer is that the primary length dependent transmission parameter for optical fibre is bandwidth and it is very difficult to measure accurately for installed cabling. As a result the application standards bodies such as IEEE had never specified a requirement – instead they had specified how far the application would operate over an optical fibre of a known bandwidth. So OF-300 applications were those that were stated as being supportable over 300 metres. Equally importantly, because the applications standards bodies had not defined bandwidth requirements, the Class of a given application had to be product specific – so 1000BASE-SX was a Class OF-500 application using OM2 50/125 mm optical fibre but was in the OF-500 group for the OM1 equivalent. This is wholly contrary to the copper concept where the Class is component independent.

The second intention was for each Class to define a performance envelope to assess compliance.

Copper channels are assessed against a limit for each of the necessary transmission parameters e.g. a maximum attenuation 24 dB at 100MHz for a Class D channel. However, optical fibre channels are required to be design-compliant i.e. the limit depends on the length and the number of splices and connections within the channel. So although the performance envelope for an OF-500 channel is 3,25 dB, a channel of length 250 m would not be assessed against this value – one would have to calculate the limit based on the length. The reason for this is to avoid the use of non-compliant components but this resulted in most designers ignoring the optical fibre Class concept completely because their implementation and acceptance rules had not changed at all.

The final nail-in-the-coffin has been the rapid reduction of, and the multiplicity, of maximum channel lengths using different optical fibre types.

So, unless comes up with a bright idea, it has been decided to remove the concept of optical fibre cabling Classes but to retain the requirement that “the attenuation of channels at a specified wavelength shall not exceed the sum of the specified attenuation values for the components at that wavelength (where the attenuation of a length of optical fibre cable is calculated from its attenuation coefficient multiplied by its length)”. It is hoped that this will be the end of an unhelpful diversion in the life of the generic cabling standards.