|
The Future of Structured Cabling
Many IT managers are unsure of which new technologies to follow. This is partly due to the considerable marketing hype from the manufacturers. The standards bodies are also taking time to ratify category 6 and this is causing uncertainty within the industry. We will be presenting a strategy and method for structured cabling that will be the most cost effective solution for today's standards. We will examine:- The Implications for future network design What is likely to be technically achieved. What will we require in the cabling infrastructure. Summary The advantages offered by high specification copper cabling (category 6 or 7) in data networks are likely to be limited and short lived. In order to avoid serious network disruptions in the future and save costly reworking, "optical fibre to the desktop" deserves serious consideration for new projects. This is also a business critical decision which could impact on your competitiveness. We should do what we can now to ensure our new network installations do not become the legacy networks of the (perhaps none too distant) future. Monotonous Rise in Network Data Rates since the 1980's when network speeds were dictated by the rotor speed of mechanical mechanisms (clunkety-clunk teletypes), the networks have seen a monotonous rise in data rates. The rise has followed "Moore's Law" closely in the past but recently, it has shown evidence of an acceleration faster than that predicted by the law. One-gigabit data rates are available now and 10-gigabit rates are predicted by 2005.
Vertical LAN In any building distribution, the "vertical" network carries about 80% of the traffic. This part of a building local area network (LAN) is business critical but fortunately, it is relatively easy to make changes to it. Quite often, this part of the network is realised as optical fibre. At one time, the bandwidth offered by optical fibre was regarded as unlimited but that is not the case now. For example, at 1-gigabit rates, there is a length limit of about 220m when using popular 62.5/125µm multi-mode fibre. This can be increased to about 500m when using 50/125µm fibre. Greater distances imply the use of monomode fibre. Any changes to the building vertical LAN can usually be made in parallel with the existing network services so disruption of the service to the desktop can be kept to a minimum. Because of this, costs to the organisation due to interruption of network service can be kept small. Horizontal LAN When considering the "horizontal" distribution within a building, we see the same increase in data rates as predicted by "Moore's Law", showing the recent accelerating trend but lagging the vertical data rates by 5 years.
So, we would expect 1-gigabit data rates by year 2000 and 10-gigabit data rates by year 2007. Indeed, 1-gigabit network interface cards (NIC) are being shipped now; it is just a matter of time before we start receiving demands for gigabit networking at the desktop. High bandwidth/critical bit rate applications such as VOIP, video streaming, video conferencing and computer aided design (CAD), are becoming common. Unlike the building's vertical LAN, the horizontal network is very costly to upgrade. Installations always mean serious disruption to normal work for a building's occupants, so this sort of activity is normally undertaken when there is a major refurbishment of the fabric. Cableways have to be formed in, on or under walls, floors and ceilings. The installers have to ensure that results are not unsightly but inevitably there is damage to the amenity value of the rooms through which cables have to pass. Care has to be taken to provide adequate and presentable cable containment for both power and voice/data cables. This all means expense. The cost of ownership is high. It was precisely because of these high costs that unshielded twisted pair (UTP) ethernet was developed. This permitted the use of existing, low grade, category 3 telephone cabling within buildings. Strenuous efforts have been made to improve the bandwidth capabilities of this popular medium. However, once firmly established, economic improvements to this system could be made only by redesign of the cable, leading to the production of category 5, 5e and 6 cables with category 7 in the pipeline. These improvements have enabled the development of 10-megabit, 100-megabit and 1-gigabit ethernet over UTP. Indeed, it has been said that you can transmit and receive data at gigabit per second rates with acceptable error rates over barbed wire. It's only a matter of how much you are willing to pay for the encoder and decoder! Limitations of Copper Cabling There are intrinsic limitations to the use of UTP copper cabling at very high data rates. These are: The susceptibility of the metallic conductor's properties to fluctuations in temperature.The increasing loss of signal strength at high data rates due to its radiation from the cable. The relatively large effects of external interference on weak signals, particularly near the ends of the cable. This is especially important where data cables share adjacent cable containment with mains wiring. The increased signal coupling between pairs of conductors in the same cable (crosstalk). This situation is exacerbated at the ends of the cable where it is attached to the hardware. At the connector, relatively high power transmit signals are launched into the cable in very close proximity to the receive signals that have already been attenuated by radiation loss along the cable's length. These weak receive signals can be swamped by the unwanted coupled signal from the transmitter. To overcome these difficulties, more has to be spent in the design of the transceiver hardware. The cost of these items will rise as bandwidth is increased. Category 6 cable has been developed in order to address some of these problems but, at the moment, no standards have been ratified. Apart from these "fluid" aspects of copper solutions, there are other factors regarding their choice that should be borne in mind. Perhaps the most important consideration is the rather low bandwidth improvement that category 6 cable will offer. The ever rising expectations of network bandwidth will soon overtake the 2.5 Gbs-1 theoretical through put that it may provide. The respite that it may give is likely to be short. There are further worries about its interoperability and its backward compatibility with existing installations. Category 7 cable is still under development and many experts in the industry believe the R&D to be vapourware. Worst case predictions for category 7 at 10-gigabit speeds imply a length restriction of about 25m. This would be a severe limitation. Any improvement will be at the expense of costlier transceivers. In any event, these distance constraints at higher data rates are onerous for UTP cabling. Category 7 then, must be at least a fibre solution? Optical Technology Optical technology, on the other hand, is already advanced. It has been improved by the development of extended bandwidth multi-mode fibre and by the development of the VCSEL laser. The refractive index profile of ordinary multi-mode fibre does not follow the ideal parabolic variation across the diameter of the core. These fibres have a small index discontinuity located at the axis of the fibre that prevents paraxial rays from being transmitted efficiently. These are the rays that suffer least dispersion and it is low dispersion that permits higher signalling rates to be used. Extended bandwidth multi-mode fibres have been developed with an improved refractive index profile. They have increased the distance over which gigabit ethernet may be transmitted by multi-mode fibres to about 1km. The VCSEL laser is capable of launching paraxial rays along the centre of an optical fibre and is the ideal complement to the use with extended bandwidth fibre.
Traditionally, copper networks have been cheaper than their optical counterparts, but as data rates increase, the costs of installing these media will converge. This is likely to happen at rates of about 2.5 Gbs. The Future Obviously, all predictions about future requirements are fraught with difficulty and we must rely heavily on past experience. Errors of judgement when dealing with the horizontal networking are expensive and have long lasting consequences. There is usually only one chance to get it right as the lifetime for horizontal wiring is 10 years or more. By way of contrast, the lifetime of a p.c. is about 2 - 3 years and that of an ether switch 3 - 5 years. In view of this, it is crucial that the right choices are made for horizontal cabling.
For 10 Mbs working, copper category 3 cabling is quite adequate. For 100 Mbs working, copper category 5 cabling is required. For 1 Gbs working, copper category 5e cabling is required. For 10 Gbs working, extended working 50/125µm fibre optic cabling is required If, as trends suggest, we may expect 10 Gbs operation within the lifetime of the wiring and in view of the fact that horizontal wiring is expensive to alter, we should try to make any current installation as so called "future proof" as we can. As the solution offered by improved copper cabling is likely to be transient and is likely to be just as expensive as fibre, we need to consider the provision of fibre for all new work. If, because of current cost, fibre is not actually installed in any given project, at the very least, space provision in the cable containment ought to be made in order to accommodate its future adoption. One way of achieving this without laying any actual fibre is to make use of "blown fibre" conduits. This is a technique whereby conduits for fibres are installed alongside traditional wiring. They may subsequently be brought into operation by using compressed air to blow fibres down the pre-installed conduits to their destination. This has the following advantages: There is minimal disturbance to decoration. There is minimal disruption to the normal working patterns of the building's staff. The extra bandwidth that fibre offers can be provided on a piecemeal basis and in parallel with existing wiring when need or budget permits. It allows expenditure to be deferred. Blown fibre is "raw" fibre (about 0.1mm diameter). It is protected from damage by the pre-installed conduits. It is cheaper than traditional fibre cable where fibres are encased in the familiar plastic sheath. The grade of fibre can be changed if circumstances change or if damage occurs. Old fibres can be quickly and easily withdrawn and replaced by different ones. Conduits are small and can be carried by traditional cable containment. Damaged conduit can be repaired by cutting out and the damaged section and replacing it with new tube. As fibre has all the advantages of proven technology, far superior immunity to electrical noise, intrinsic electrical safety, superior bandwidth capability, a high degree of data security and the ability to carry signals over long distances permitting the use of "collapsed backbone" topologies, its use or potential use ought to be seriously considered for all new cable installations. If history teaches us anything, it's that no solution is "future-proof"
|
