Linking the numbers
Structured cabling is covered by Standards that seem to get a new number every time you look. George Georgevits offers a useful guide.
In 1989 a client company asked me to verify the performance of some Category 3 product it was thinking of selling.
This was my first involvement in testing balanced-pair generic structured cabling systems.
The only available Standard was an early draft version of the first edition of EIA-568. Fortunately, I was able to set up rather limited test equipment to the required configurations and perform the necessary insertion loss and crosstalk tests.
In so doing, I developed an interest in the field.
As time passed, industry adopted the concept of generic structured cabling systems for buildings. Things progressed rapidly, to Cat 4; Cat 5, Cat 5e; Cat 6, 6A; Cat 7, 7A; and now Cat 8. How things have changed.
People regularly ask questions about the capabilities and limitations of these grades of structured cabling infrastructure and how to test for Standards compliance.
All twisted-pair structured cabling systems are made up of links and channels.
Links and channels are made up of components, examples of which include cables, outlets, patch cords and fly leads, and cross-connects.
The performance requirements for various links, channels and components are specified in the structured cabling system Standards.
There are two international Standards that specify the various grades of generic structured cabling systems: EIA-568-C series (US) and ISO 11801 (European). Each of these also references many other normative, and also informative, Standards.
The terminology and the performance requirements of the ISO and EIA Standards are similar, but not identical.
In Australia, AS3080 is the relevant Standard, this being a close adaptation of ISO 11801.
EIA and ISO, plus their normative references, fully specify the requirements for various performance levels and compliance testing – in the field and in the laboratory.
EIA was responsible for initial development of the component performance and component testing requirements, with an emphasis on unshielded twisted-pair systems.
ISO approaches the subject more from a link and channel performance viewpoint, with special emphasis on higher performance shielded twisted-pair systems.
All of these grades of structured cabling systems use 100Ω balanced twisted-pair cabling. Many categories come in unshielded and shielded variants.
The transmission performance of each category has been designed to support the spectral requirements of signals associated with a particular data transmission protocol.
Support for the protocol by its related data cabling system class is guaranteed by the Standards to a reach of 100m, unless otherwise specified.
This harmonisation between cabling hardware and data transmission protocols was achieved in each case by the cabling Standards bodies liaising with the corresponding protocol development Standards body (the IEEE 802.3xx series Ethernet Standards committees).
As a general rule, transmission performance improves with category number. The suffixes ‘e’ (enhanced) and A (augmented) refer to improved versions of that particular category.
The major improvement in performance with each category number has been achieved by:
- increasing the types of transmission performance tests required; and
- tightening the performance requirements of existing tests used for qualifying lower categories.
Link and channel performance is tested in the field with specially designed testers. However, field testing assumes that the individual link and channel components (cable, patch cords, outlets, etc) have already passed the relevant component-level laboratory tests.
A weakness with this approach is that with some non-compliant components (eg: poor-quality jacks) the field tester can show a ‘pass’ for a link or channel test, but the link or channel may not support the required protocol (eg: Gigabit Ethernet for Cat 5e systems).
This is because the performance of the cable tends to dominate the link performance. Reflections and/or crosstalk induced by low-performance jacks may not be enough to fail the link test but can prevent proper functioning of the link.
Capabilities and limitations
First, let us consider existing technology systems.
Cabling systems using cabling below EIA Cat 5e (or ISO Cat 5 equivalent) are restricted to legacy systems operating at link speeds of up to 100Mbps. These are generally no longer found in larger commercial premises. The ever-increasing demand for bandwidth is rapidly making them obsolete.
The bulk of installed cabling is Cat 5e running Gigabit Ethernet. Such an arrangement is robust, cheap, highly reliable, easy to maintain, easily expandable and supported by a large number of vendors of cabling hardware and terminating equipment.
The 1Gbps speed is more than adequate for all but the most demanding commercial applications.
Cat 6 is the preferred choice for new cabling systems and upgrades. The overall installation cost is only marginally more than for Cat 5e, if we include labour costs.
Such systems can support 10G for short links (up to 55m), provided they are qualified to TSB155 requirements at the time of installation.
Given that the average horizontal cable run in a typical office is somewhere around 40m, Cat 6 provides a cheap alternative to Cat 6A in many situations.
Cat 6A will support 10G to a full 90m horizontal cable run plus 5m of patch cord/fly lead. It is the cabling system of the future, and costs are rapidly coming down as production ramps up.
It is still more expensive than Cat 6, but if the plan is to occupy the premises for many years to come, and there are many long runs, then Cat 6A is the logical choice.
Now to consider Cat 7 cabling and above.
Cat 7 was introduced by ISO over ten years ago. It use shielded cabling and a non-RJ45 shielded connector. Consequently, it has always been expensive to purchase and difficult (and thus expensive) to install.
In 2009, the EIA ratified the Cat 6A Standard. This offered similar performance for a fraction of the price, using unshielded cable and conventional RJ45 connectors.
Consequently, Cat 7 was never adopted by industry, as it was not competitive in price and did not offer any significant performance advantage over Cat 6A.
Similarly, when Cat 7 was enhanced to Cat 7A to further improve its performance, nothing had changed. It was still expensive, and it used a non-RJ45 connector. Furthermore, there was no protocol that required the increased performance offered by Cat 7A.
In my view, Cat 7 and Cat 7A were total failures, insofar as neither was adopted by industry. I am not aware of any terminating equipment equipped with Cat 7/7A connectors.
In June 2016, the EIA published Standard ANSI/TIA-568-C.2-1. This defines the performance requirements for Cat 8 cabling.
In brief, Cat 8 has been designed to support a specific data centre application – a cheap, short distance, high-speed link between servers and switches.
The EIA recognises only one version of Cat 8, using 22AWG shielded cable and a shielded RJ45 connector. It uses only two pairs and is designed to support 25/40Gbps for runs between 5m and 30m.
ISO recognises a second version with a non-RJ45 connector. My guess is that this will go the same way as all other Standards that used non-RJ45 connectors.
At the time of writing, Cat 8 product has yet to reach the market, but unless you are involved in installing a data centre, this will not be of concern.
Last year, the IEEE-SA Standards Board approved IEEE 802.3bz-2016. This was written to define 2.5GBase-T and 5GBase-T running over Cat 5e and Cat 6 cabling for runs up to 100m.
This is intended to be a lower-speed, long-distance alternative to the short-distance 10G defined in TSB155. It provides a way of squeezing the last drop out of existing Cat 5e and Cat 6 installations. Time will tell whether this is worth considering.