By Dustin Guttadauro, L-com Global Connectivity
Although other technologies often take center stage in the Ethernet “ecosystem,” the cable types and connectors required to implement Ethernet’s increasing speeds are every bit as important-and often confusing. When viewed over the long term, those with deep pockets might consider an enterprise-wide investment in either singlemode or multimode fiber as the obvious choice. But fiber is not a panacea and as this article will illustrate many organizations, or at least large portions of them, can be served quite well by Category 6A twisted-pair copper cables.
Category 6A Ethernet cables have been available since before 2008 when the ANSI/TIA-568-B.2-10 Category 6A standard was released. However, their adoption began slowly as nearly 9 years ago, most organizations had no need for them as data rates were well below those of today, which are typically at least 1 Gbit/sec and rapidly increasing. Data centers were the primary areas where Category 6A Ethernet cables were initially used, so it’s not surprising that this remains their largest application.
Category 6A cables have evolved considerably since their introduction, when they were 50 percent larger than their predecessors. Today’s small-diameter Category 6A cables are about 15 percent larger, which minimizes considerations of added size and weight as well as how many cables can be placed into a cable tray or conduit. They also have much better alien crosstalk performance and noise immunity.
These more-advanced Category 6A cables and their small-diameter variants are being adopted by greater numbers of organizations that recognize the need to update their current Ethernet connectivity solutions to support applications that frequently transfer very large files to and from the data center. As a result, Category 6A cabling is being deployed in many more areas of the enterprise and even for backhaul of cellular and WiFi traffic, an application that will grow in importance as the fifth generation of cellular networks with data rates above 1 Gbit/sec are rolled out early in the next decade.
10GBASE-T is soon likely to be the choice for most organizations, driven by data center backbone speeds moving to 25, 40 and eventually 100 Gbits/sec. As data rates increase so too will the number of organizations switching to optical fiber, at least in those areas where its vast bandwidth and immunity to EMI (electromagnetic interference) and RFI (radio frequency interference) make it extremely appealing. In the meantime, copper solutions will continue to improve so it’s likely that both technologies will coexist for many years to come. One of the greatest reasons for this is that copper-based systems cost less, are much more familiar to installers, and use the simple RJ45 interface that is almost universally used in all generations of Ethernet. They also support Power over Ethernet (PoE) at increasing power levels, currently up to 100 W, thanks to design techniques that dissipate heat around the circumference of the cable to eliminate areas of high temperatures (hot spots).
The first step in making an upgrade from existing Category 5e (or earlier) cabling is evaluating the needs of where it will be implemented. For example, most organizations have what could be called “standard,” “enhanced,” and “performance-centric” centers that rely on Ethernet connectivity. Standard-class operations, which usually encompass most departments in an organization, have minimal speed requirements: 1 Gbit/sec is more than adequate.
The enhanced category includes areas in which data must be transferred to and from the data center very quickly, and file sizes are often enormous. These can include engineering, a creative team whose work includes both high-resolution graphics and video, and basically anywhere high speed is essential to productivity. In these cases, data rates of 1 Gbit/sec should be acceptable, but up to 10 Gbits/sec might be preferable. At the high end is the data center itself, which requires the highest possible throughout and 10 Gbits/sec is essential.
Standard-class operations today typically use Category 5e or Category 6 cables, which in terms of throughput are fine for now and in the future. The enhanced class of operations requires Category 6A cable as it satisfies speed requirements and is a newer, more-advanced standard.
The data center represents the area in which there are multiple cabling possibilities. In contrast with other corporate groups in which users are typically widely distributed, data centers are designed specifically to support data processing, storage, and distribution, with subfloors and overhead areas. When compared with the space constrictions and other issues central to upgrading to newer cabling standards in other areas they are a comparative walk in the park.
Intra-data center cabling also requires shorter cable runs, so the distance limitation on Category 6A of 100 meters (including 10 meters of patch cords) while supporting 10-Gbit/sec data rates is not an issue. In addition, as the data center is the “mother ship,” which all things data pass through, and requires comparatively less cable, it is economically well suited for the use of fiber and its essentially unlimited bandwidth. The overall cost of using fiber today is determined not by the cost of the cable itself (which continues to decline) but by the hardware associated with it, such as transceivers and switches that remain expensive. If the goal is to become “futureproof,” the additional cost of fiber can be amortized over years of Ethernet enhancements with relatively low upgrade cost.
Having delineated the three major categories in an organization, their needs and thus cabling requirements, the next issue is how to implement Ethernet upgrades within them. In new buildings this is far less difficult as they are typically designed and constructed with future requirements in mind. However, for existing structures that currently use Category 6 or earlier cables, other things must be considered-foremost of which is space, or the lack of it.
To meet the more stringent needs of Category 6A such as 10-Gbit/sec data rates (10GBASE-T), an operating frequency of 500 MHz, and a maximum theoretical distance of 100 meters, significant changes to cable design were required. They include the need to more-effectively address noise and alien crosstalk (noise between cables in a bundle), grounding, bend radius, the demands of PoE, and overall construction. The result is that these cables are larger than their predecessors, and while cable manufacturers have minimized the increase in size, this can still present challenges, especially in existing structures.
Upgrading to Category 6A from Category 6 or earlier Ethernet cables places demands on the space limitations imposed by conduit, cable trays, and all areas of a building through which cables pass. The result may be the inability to use the same number of cables as the existing solutions. Solving this problem typically requires additional cable routing hardware, which can be expensive and why it’s extremely important to understand where Category 6A cables are a necessity.
Recognizing the issues surrounding the increased size of Category 6A cables, some manufacturers, including L-com, have created cables using 28-AWG copper wire, which is thinner than the traditional 23- and 24-AWG wire used in Category 6 and other Ethernet cables and results in cables whose outside diameter is reduced by half. They are shielded to fend off EMI and crosstalk, support PoE at reduced power levels, maintain 10-Gbit/sec performance, have a tighter bend radius than standard Category 6 cables, and are more flexible, making them easier to work with.
Shielding and other considerations
Any type of shielding provides some level of isolation from signals emitted by the host system, but the fact that a cable is stated as being shielded is no guarantee that it will be effective. This is because its shielding performance depends on factors including how well it is constructed, the materials from which it is made, grounding, and the effectiveness of the Faraday cage created by the shielding.
EMI and RFI are increasing concerns as Ethernet cables are often co-located with systems that produce RF energy either as a primary function or incidentally as a byproduct of the frequencies at which they operate. The problem has also long been associated with the proximity of Ethernet cables to electrical cables, as they can include 50- or 60-Hz currents and noise spikes from the electrical cable to the Ethernet cable.
The solution is the same as always but more important with Category 6A: Keep cables as far as possible from power cables as well as the sources of line-frequency interference such as fluorescent lights, some medical equipment, motors, air conditioners, and other sources of low-frequency energy.
There are other considerations when installing Category 6A cables, and one often missed is achievable distance. Although the standard specifies maximum cable length of 100 meters, it includes the distance covered by patch cables. This means that the total specified cable length must include not just the horizontal run but the patch cables on each end as well.
Another factor to consider is that electrical conduit and termination boxes aren’t well-suited for use with Category 6A cables because their allowable bend radius is less than the cables can reliably achieve. Attempting to circumvent this rule will result in performance degradation as well as possibly failure caused by kinking and stress.
The environments thus far described are typical of enterprise environments. However, factories, oil and natural gas refineries, and other industrial applications represent a significant portion of Ethernet cable installations. In these scenarios, repeated flexing, vibration, shock, crushing, temperature cycling, electrostatic discharge and intense magnetic fields, and exposure to salt and corrosive chemicals are commonplace. Cable manufacturers offer jacketing material to meet these needs that is often made from polyurethane or FR-TPE (flame-retardant thermoplastic elastomer) that is extremely resistant to abrasion, chemicals, sunlight, and water. Other jacketing is also available from those manufacturers offering customization.
The quality of cables and connectors is, or should be, important to anyone who buys them, as reliability, performance, and many other factors vary among manufacturers. Consequently, it is important to choose manufacturers with an established long-term record of accomplishment and stand behind their products and specifications.
It is also common these days to find counterfeit products in the marketplace. Detecting counterfeit cables requires the buyer to thoroughly examine them to determine whether they have the stated-diameter wire, that their twist ratios are consistent, and their sheathing is well constructed (or even present), among other things. For example, it should never be taken for granted that if the cable is promoted and labeled as using 24-AWG wire that this is what’s inside. Close inspection can reveal that the cable actually uses 26-AWG wire and may be copper-clad aluminum (CCA) rather than copper. CCA reduces performance and reliability.
Connector quality and its interface to the cable should also not be assumed. For example, cheap or counterfeit connectors may have a flash layer of gold on their contacts rather than 30 or 50 μm gold plating. The flash gold layer will rapidly wear off in test and measurement and other applications in which cables are repeatedly connected and disconnected.
Industry pundits have for many years projected the death of copper-based Ethernet cabling only to repeatedly be proven wrong, as Category 6A cables amply demonstrate. There is no question that optical fiber has immense benefits in some situations where copper-based solutions cannot compete, and this is not likely to change soon. That said, most Ethernet applications are well served by copper-based solutions, and this is not likely to change soon either.
Dustin Guttadauro is a product manager with L-com Global Connectivity (www.l-com.com).