Category 8 updated

A future proof structured cabling system uses both copper and fiber optic cabling. The reason is that copper has shown to be the much more universal medium, new applications in past years such as VoIP or PoE demand copper, with fiber you already have excluded such applications. As we do not know what the future will bring, it surely makes sense to continue to deploy copper as an alternative. So in case fiber will not be able to handle an application, there will be a copper cable to shoulder it. In an ideal world, both mediums a used in parallel for data transmission. If one medium fails, the other one can jump in. Of course in order to have real redundancy, it would be ideal if new standards were equal for copper and fiber. However the fact is that they rarely come out in parallel. Fiber usually comes a bit later and then offers much better bandwidths. So users in data centers have been hooked on fiber transmissions failing to look at the copper side. Ideally both systems work simultaneuosly, error rates can be reduced as the best of both solutions is selected.

However we do not live in an ideal world. Unfortunately the race between fiber optics and copper is as fierce as ever. And it continues to look like the race between hedgehog and rabbit. In 2006 for example IEEE standard 802.3an came out for 10 GB/s over copper, shortly after IEEE 802.3 ba came out for fiber optics, and once again the standard for fiber was much faster, it was 40 and100 Gb/s already ten years ago, so fiber had the upper hand.

History repeats itself. Ten years later in July 2016 in San Diego IEEE met again and for copper we now have 802.3bq for 25GBase-T and 40 Gbase-T, while fiber once again is ahead with 100GB/S and 400 GB/s. When we look closer at the copper standard, we see that we will use a 4 pair symmetrical and shielded copper cable system over the length of 30 meters. How this is done is described in full detail over 220 narrowly printed pages. There is also a technical report of ISO/IEC 11801-9901 ?Guidance for balanced cabling in support of at least40 Gb/s data transmission“ which is an international proof of concept explaining how things should be done. And once again we have fiber running after copper and having a head start as the fiber standard IEEE 802.bs (which will probably come out in 2017) will go up to 400 GB/s.

In 802.3 ba copper had the additional difficulty that it specified eight pair Twinax cables, which were limited to seven meters, CR4, in the top of rack configuration. The eight fibers of OM3/OM4 which were used for end of row configuration could reach out over the full 100 meters. In the new copper standard port to port links with patch cables with 5 or 10 meters are used and server switch links can go up to 30 meters with two connectors at the end. This set up is for full duplex. For 25G/s it will be 1250 MhZ versus the 2000 MhZ in the 40 G/s bandwidth range.

Unfortunately another complication, which we used to have in the past continues unchanged. We will have a different standard on both sides of the Atlantic. It is good news that the new copper cabling is not as confusing as in the past ( we all remember the discussions about the difference between a 6A in the US and a 6A in Europe, yes, it is absurd, the two standards were differing in the way that the 6A in Europe was written with an A that was written lower on the line, no joke.) Fortunately the San Diego crowd spared us this time such confusion that drove to dispair countless project teams and uselessly ruined many projects as it came to misunderstandings all the time. In 2016 we have become smarter. Yes, there is a split between the two standards for copper, there will be a 8.1 in the US and a 8.2 in Europe. Now at least that will not be as easy to confuse as the horrible 6A (US) and 6A (written low in Europe) chasm of the past. Fortunately for so many projects, it will be much easier to distinguish between the two.

Still the discussions about which standard is better are fierce. In Germany for example the engineering body VDE allows only 8.2, arguing that the American way of doing things will be more wasteful. Both NEXT and FEXT in the US will be 20 dB higher with the 8.1 versus the 8.2 standard. A lot of electricity will therefore be saved in Europe on the 40 GB/s transceivers alone, simply by using 8.2 instead of 8.1. Saving electricity is always a good argument in Germany.

The other reason why of course Germany especially is unhappy to use an American system called 8.1. is, because it is not backward compatible to Cat 7 and Cat 7A cabling. A 8.1 only sees Cat 6A cables, which of course is unacceptable in Germany, where 90% of all installations are done in Cat 7 or Cat 7A. In the US this never mattered because Cat 7 is practically not used. But users in Germany would be forced to rip out most of their old cabling if they accepted 8.1.

BSRIA offers an interesting statistics, saying that globally there are 10 or 20 million Cat 7 and Cat 7A cable installations. Now the interesting question will be how the dice will be cast this time. It could well be that for the first time in IT history, the European way of cabling will be predominant in the world. Already in other ways the US is loosing ground. The mere fact that Shielding will be the norm in both 8.1 and 8.2 points to the fact that technically what was in the past considered pure over-engineering by over celoius teutonic Krauts, has now become a sort of preferred way of doing things in copper and in many regions of the world. We can see this trend clearly in Eastern Europe. This region was due to Russian influence firmly in the UTP field, but entire countries like Poland or Slovakia are abandoninning the unshielded camp and now are predominantly doing German style FTP cabling. 

In order to summarize this blog. Ideally 8.2 will be the winner in the copper race and in the copper versus fiber optics race there will be no winner at all but coexistence of both systems.