Is it a round peg in a square hole, or a perfect fit?
BY Dr. Andrey Semenov, IT Co. and Igor G. Smirnov, Signamax Inc.
The specification and selection of information technology (IT) systems and services often includes a presumption that structured cabling will serve as the foundation of the IT infrastructure. Although an IT cabling infrastructure foundation is a given, other types of IT systems (e.g. wireless networks and data over power cabling) are generally considered to be niche solutions and typically do not play a defining role in the physical layer infrastructure planning.
Structured cabling systems feature a wide range of supported networking equipment and associated transmission protocols. Leading IT equipment manufacturers are able to offer their customers a variety of products of the same functionality yet featuring different technical transmission parameters. That variety presents a designer of structured cabling with a challenge to choose an IT cabling solution that satisfies present and future needs of a particular project.
Balanced twisted-pair cabling has traditionally been classified by various categories of transmission performance. The higher the category, the better the electrical performance and the higher the maximum rate of information transmission. One tradeoff to higher performance, however, is that higher transmission performance results in higher component acquisition expenses as well as increases in cabling installation and field-test complexity, which significantly affects the overall project cost.
|As link lengths grow, the disparity between Category 6 and other media–Category 6A, Category 7 and optical fiber–widens.|
In general, the task of choosing the appropriate category of structured cabling components may become easier if a telecommunications application is used as a reference point that serves as the focus application domain of a certain type of cabling. For the sake of this article, we will focus on the decision matrix leading up to Category 6A cabling, which represents the most interesting case because it can be considered the most advanced type of structured cabling being implemented on a broad scale.
Category 6A backstory
Category 6A cabling standards emerged in order to support information transmission rates of 10 Gbits/sec. Prior to developing Category 6A cabling components, manufacturers of Category 6 cabling initially focused their attention on redefining the frequency range of interest for Category 6 from a ceiling of 250 MHz to a ceiling of 625 MHz. After years of research and study, it was determined that a frequency range of 1 to 500 MHz would be required to support the 10GBase-T application.
Perhaps the most significant technological hurdle to overcome would be the characterization of alien crosstalk. Ironically Category 7 cabling, which evolved in the mid-1990s, fully supports the 10GBase-T alien crosstalk requirements and did not require any modification in order to support the 10GBase-T application. Category 7 cabling, however, was not widely accepted at the time the Institute of Electrical and Electronics Engineers (IEEE; www.ieee.org) was developing the 10GBase-T standard. Plus, Category 7's higher performance is accompanied by the aforementioned higher pricetag, and Category 7 could not ensure cost parameters required by practical applications used in common general-purpose networks. For these reasons, Category 6A cabling standards have emerged as the predominant cabling category in support of information transmission rates of 10 Gbits/sec.
Regarding the emergence of Category 7 cabling in the mid- to late-1990s, these components were developed primarily to support 622-Mbit/sec applications such as Asynchronous Transfer Mode (ATM). ATM-622 did not use the parallel transmission techniques used by Ethernet applications and its performance is based on binary encoding of the line signal. When ATM-622 was developed there was no task to match active equipment transceiver parameters to those of the passive part of the communications channel. The lack of optimization led to a situation in which a Category 7 four-connector, 100-meter channel was used for a little more than 10 percent of the its theoretical 55-Gbit/sec channel capacity as calculated according to Shannon's Theory. With such a low-percentage use of Category 7's functional capabilities, it is not surprising that Category 7 does not measure up as an economically advantageous option for 10-Gbit/sec transmission.
From the data presented in the chart on page 5 and the considerations already made in this article, we can draw a couple conclusions.
Category 6A technology more precisely matches the physical application of structured cabling from the system-level point of view.
For relatively short link lengths and transmission rates up to 10 Gbits/sec, balanced twisted-pair cabling and components are a preferable option for system users.
|While permanent-link lengths in commercial buildings average around 40 meters, average links are significantly shorter in data centers.|
To comment briefly on the concept of "matching the field of application," Category 6A was developed with the objective of supporting 10-Gbit/sec Ethernet signal rates up to distances of 100 meters. The Shannon capacity of a Category 6A channel is 18 Gbits/sec, while some advanced contemporary systems using screened component designs increase the Shannon capacity values to between 30 and 40 Gbits/sec. Either cabling setup both guarantees the normal functioning of the network interface and ensures good cost parameters of the cabling infrastructure solution as a whole. The most significant factor that defined the application area of Category 7 solutions was that the technology did not include an unscreened option.
The 10-Gbit/sec Ethernet network interfaces use the potential capacity of contemporary cabling channels to the highest degree. This achievement was made possible by the application of a complex set of technical methods. Among them are the multilevel encoding intended to minimize the bandwidth of communications channels, the correcting and compensating mechanisms used for separation of valid signal and noise–known as noise-cancelling techniques–and some additional and similar techniques. However, the tenfold increase in the transmission rate of the 10-Gbit/sec Ethernet active networking equipment compared to its precursor required more than a fourfold increase of the line-signal spectrum's upper cutoff frequency. The proportionality distortion in this case can be explained by the application of the PAM-16 code in 10-Gbit Ethernet compared to the PAM-5 used in Gigabit Ethernet.
The operation of unscreened cabling channels in extended frequency ranges up to 500 MHz leads to the appearance of significant inter-cable crosstalk, also known as alien crosstalk. Consequently, that forced the application developers once again to extend the list of transmission performance parameters required to be controlled in the field in order to commission the cabling. The new list included alien crosstalk in both pair-to-pair and power-sum versions.
Applications for 10G
The typical user of the network resources provided by an IT system is not able to adequately sense the flow of information delivered at speeds beyond several dozen Mbits/sec. This assumption is not likely to change in the near term. The fundamental requirement to increase transmission speeds up to 1 Gbit/sec, suggested just 15 years ago and made a reality at the turn of the century, has been taken off the agenda of some IT users. The reason is that videoconferencing, the most and some say the only compelling user application for 1-Gbit/sec transmission, has not obtained the anticipated broad acceptance that was forecast some time ago.
It immediately follows that the focus field of application for Category 6A technology becomes those structured cabling segments that are not principally speed-limited by human-computer interaction. Such segments naturally fit in the backbone subsystems in conventional office enterprise cabling as well as data center cabling. Therefore, we conclude that the major application for which Category 6A technology will be used on a mass scale is in data centers.
Category 6A deployment volume
The results of theoretical calculations and cabling installation practices associated with commercial office cabling systems provide estimated expenditures for the backbone implementation in structured cabling systems at 15 to 20 percent of the overall cabling cost. In the context of the IT system architecture in general, data center cabling can be considered as a remote projection of the classical structured cabling backbone subsystem. In data centers, structured cabling links are used for connection of servers to switches and mass storage devices; therefore the above-mentioned percentage may be doubled.
The means intended for acquisition of optical equipment should be excluded from the summary estimation. Manufacturers of optical-fiber components quite often bring to focus as the strong counterweight to somewhat worse economic parameters, the smaller energy consumption of network interfaces, which is critically important for data centers. Nevertheless, to use this advantage in actual practice is not possible for several reasons.
The data center structured cabling has a distinctive feature important from the viewpoint of project implementation, which, taking into consideration prevailing market rates, is the heaviest portion of the materials bill for both components and labor. The key here is the average length of the permanent link in the data center, which is less than 30 meters compared to a typical 40-meter permanent link in large commercial office buildings.
Two major factors influence the data center's shorter permanent-link length compared to that of the commercial office environment. First is how compact the data center is as an architectural facility, which is by design because of the high cost of real estate. Second is the inherent availability of a large quantity of cabling pathways under the access floor as well as overhead cabling containment systems typically located above equipment cabinets. The ready availability of cable distribution systems within data centers facilitates pathways that provide the shortest routes for cabling links.
With such short-length permanent links, the network equipment can work in short-reach mode in the majority of cases. In such situations, any optical interface loses its power-consumption advantage.
Taking stock of these considerations, we can confidently assume the quantity of balanced twisted-pair and optical fiber links in the typical data center cabling system will be approximately the same. Taking into account the higher cost of Category 6A components compared to Category 6 (about 50 percent higher on average), we obtain the total volume of the Category 6A market, by value, to be about one-third of the overall volume of the Category 5e and Category 6 office cabling systems market.
Structured cabling evolution
In the process of Category 6A technology development a number of innovative solutions were created that had no parallels in the previous generations of structured cabling. One such development was the partial-screening technique. The approach was applied to cabling products in the form of a metallic film screen with gaps in the metal coating. The gaps prevent formation of current loops without impeding normal functioning of the screen at frequencies beyond 300 MHz.
The addition of a metallic film screen to the construction of distribution cables, outlet modules, patch cords and other equipment does not worsen their mass or dimension parameters; it does, however, ensure the increased resistance to external interfering electromagnetic radiation. As a result, systems of this type increase their alien near-end and far-end crosstalk performance, in pair-to-pair and power-sum models, by approximately 10 dB.
The major advantage of the partial screening technology is that cabling may be installed per unscreened cabling installation rules, making the requirement to provide a telecommunications bonding and grounding system–which is effective at frequencies within the order of hundreds of megahertz–irrelevant.
Therefore, from a technical point of view, application of such partial screening design concepts rather effectively solves the problem of creating necessary alien crosstalk headroom, which cannot be ignored in the upper half of the 10G Ethernet frequency range. From a practical point of view, partial screening technology is advantageous because its increased crosstalk margins noticeably accelerate and simplify cabling installation.
The aforementioned shorter average permanent-link distances found in data centers has caused the emergence of a new technology that, from the perspective of its transmission performance, can be qualified unofficially as quasi-Category 6A. Cabling components that are not in full compliance with the requirements of the commercial office building and data center structured cabling standards, may in fact still support the desired 10GBase-T application. The major deficient parameters of such components are insertion loss (attenuation) and loop resistance. In spite of these performance in adequacies, cabling channels built based on quasi-Category 6A components provide the required signal transmission quality at distances up to 60 meters, which by a longshot exceeds the practical requirements of data center cabling. Aside from that, the electrical-performance characteristic loop resistance is closely associated with Power over Ethernet and Power over Ethernet Plus. In the absence of PoE or PoE Plus, loop resistance turns out to be irrelevant.
The significant gain from the quasi-6A components' use instead of the standards-based Category 6A products can be achieved because of these cables' noticeably smaller sizes; the average diameter is a little over 5mm or 0.2 inches. That makes it possible to improve balanced twisted-pair cabling mass and dimension parameters that are important for data centers, and to considerably reduce the advantage of optical-fiber links.
It should be noted that current industry standards do not consider in any manner the interrelationships between category of cabling, link/channel length, and telecommunications applications capacities.
Based on the information we have researched and presented in this article, we draw four conclusions about the deployment and use of Category 6A cabling systems. 1) Category 6A technology is a product of transmission-performance parameter optimization between cabling and active networking equipment aimed at obtaining new properties of the integrated product. 2) Category 6A technology has its own explicit application niche, which is the lower hierarchy level (i.e. horizontal cabling) within data centers. 3) Present Category 6A solution volumes should have a steady growth trend. In the near future they will be limited to about one-third of the volume of Category 5e and Category 6. 4) In the process of Category 6A development, a number of innovative solutions were created and implemented in structured cabling practices; these solutions may lead to revisions of some fundamental positions of industry standards.
DR. ANDREY SEMENOV is director of business development for IT structured cabling systems with IT Co. (www.it.ru) and head of the structured cabling systems faculty at the Moscow Technical University of Communications and Informatics. Igor G. Smirnov is product manager for Signamax Inc. (www.signamax.com).