When standards change, so do real-world test requirements
Documentation, the permanent link, and real- traffic generation are the new buzz in testing circles.
Documentation, the permanent link, and real- traffic generation are the new buzz in testing circles.
G.W. Renken / Wavetek Wandel Goltermann
In today's rapidly evolving networked environments, professional cabling installers are constantly being pressed to stay abreast of the current standards requirements, while they simultaneously look ahead to predict what will be required tomorrow-such as the migration toward Category 5E and Category 6 cabling systems and Level III testing capabilities.
Installers must carefully plan and manage their equipment investments to maintain a competitive edge, while balancing the dual-edged risks of over-investment and short-term obsolescence. In addition, they must keep in mind their corporate customers' real-world needs, including the actual network protocols and traffic levels that the installed cabling system will have to support.
Importance of standards
In a very real way, the definition and adoption of widely accepted standards represent the bedrock upon which modern interoperable networking systems are built. Without well-specified and agreed-upon standards at every level of the networking hierarchy, there would be no way to ensure networks behave and perform as required to support critical applications and communications systems.
Several standards address today's multiprotocol, multi-application network environments at all seven layers of the Open System Interconnection (OSI) model, which is used to delineate most networks. Professional cabling installers, however, have traditionally found themselves concentrating primarily on the specific cabling standards that lie beneath the OSI model's Layer 1-the physical-medium dependent (PMD) layer. The PMD layer defines transceiver technology that provides the conversion between analog signals and digital information, thereby providing the foundation that allows the higher layers to operate in a purely digital domain.
Typically, a PMD specification defines modulation, data rate, maximum acceptable bit-error rate (BER), and ambient noise, among other characteristics. When specifying the medium, a PMD standard generally can just reference applicable generic cabling standards such as TIA/EIA-568A or ISO-11801.
The relevant standards for current networking environments cannot practicably be driven by just one centralized forum, but rather are influenced by the combined interaction of many organizations and standards-making bodies. These organizations run the gamut from professional cabling and contractor industry organization to engineering forums that guide network protocols and end-user associations that deal with real-world application issues.
Throughout their evolution, cabling standards have had to constantly address the need for compatibility with existing technologies-such as connector form factors, punchdown blocks, and other termination equipment-while also embracing new, higher-speed capabilities achievable through improved unshielded twisted-pair (UTP) cabling characteristics and more accurate testing methods. In addition, the demands of different networking environments worldwide have produced two major classifications of standards: those developed by the Telecommunications Industry Association/Electronic Industries Alliance (TIA/EIA-Arlington, VA) and those developed by the International Organization for Standardization (ISO-Geneva). TIA/EIA standards have provided the primary guidelines for installing and certifying North American cabling installations, while ISO standards have been used throughout Europe.
The TIA/EIA-568A and ISO-11801 standards constitute the primarily commercial building telecommunications cabling standards that are currently finalized and approved. While the TIA/EIA Categories and ISO Classes do not have a perfect one-to-one correspondence on all details, generally they match up. TIA/EIA Category 3, 5, and 6 align with ISO Class C, D, and E, respectively. The TIA/EIA Category 5E specification does not closely align with any ISO specification.
The TIA's telecommunications systems bulletin TSB-95 has been ratified and widely accepted as the basis for Category 5 testing and certification. TSB-95 augments TSB-67 and TIA/EIA-568A by defining the additional Category 5 measurement parameters of return loss and equal-level far-end crosstalk (ELFEXT). The new measurements of return loss and ELFEXT were incorporated at the request of the Institute of Electrical and Electronics Engineers (IEEE-New York City) 802.3ab committee, which is responsible for defining the standards for transmitting 1000Base-T-Gigabit Ethernet over copper wiring. These more stringent measures were needed because 1000Base-T requires a multi-transmit environment in which all four wire pairs transmit in both directions simultaneously.
Currently, TIA/EIA-568A Draft 12 represents a proposed revision and update of the entire TIA/EIA-568A specification that incorporates all changes to date. It also defines a new Category 6 that will essentially be the equivalent of the ISO's Class E. Category 6 will be the nomenclature applied to cabling systems using 8-pin modular connector styles and certified to carry 200-MHz traffic. Category 6 will also require testing to newly specified Level III accuracy, which will incorporate all the existing tests used in TSB-95 at the Category 5E level. The movement from Category 5 at 100 MHz to Category 6 at 200 MHz of usable bandwidth (and 250 MHz for the test suite) requires a significant improvement by as much as 10 dB in each of the critical radio-frequency parameters that characterize a test device's accuracy.
Two key parameters
Since the initial establishment of broadly accepted industry standards, they have primarily focused on clearly defining two key parameters: performance characteristics of components, including cable and connecting hardware, and the installed link's transmission capabilities.
The development of standards has also provided several benefits within increasingly complex network environments, including consistency of cabling design and installation, conformance to physical and transmission-line requirements, a basis for examining a proposed system expansion and other changes, a consistent structure for uniform documentation of cabling installations, and improved interoperability across mixed-vendor environments.
In terms of capital outlay, documentation is the least expensive network-troubleshooting tool. However, it can be the most powerful resource available to local-area-network (LAN) service personnel when the system is down. A network's inherent maintainability is only as healthy as the effort put into its documentation plan. For network administrators, one of the worst possible situations occurs when a simple upgrade or replacement turns into a staff "resource sink" because of a lack of information. In such situations, getting a device to function properly often requires effort that would otherwise be unnecessary. By meticulously documenting each network link and device upon installation or upgrade completion, users can keep the network highly maintainable.
A professional data-communications cabling installer should provide the customer with complete documentation of the network topology, type or types of cabling used, and a wiring map showing all crossconnects and end points. The installer should also label each cable in a way that is meaningful to both the installer and client. The installer should also record the location of hubs, switches, concentrators, repeaters, patch panels, and any other active or passive interconnection hardware. Cable-certification printouts should show all relevant parameters for the specified Category, such as length, impedance, connectivity, attenuation, and near-end crosstalk for all cabling components in a link. For example, the documentation must specify such details as the maximum number of nodes and the limitations of cable length.
It is important to clearly delineate the network-design limitations to provide a well-established baseline for both optimizing network performance and implementing future moves, adds, changes, and upgrades. In these days of heterogeneous service convergence and mixed-media networks, it is far easier to clearly document topological design rules, limitations, and unique variations rather than expect to remember them when future changes or problems arise.
Basic link, channel, and permanent link
Beginning with the TIA/EIA-568A standard, international cabling standards have specified a hierarchical infrastructure with individual desktop connections star-wired back to telecommunications closets (TCs) via horizontal links of no more than 100 meters. These horizontal cabling links have traditionally been tested either as basic links or channel links.
According to the standards, a field-tester-supplied test cord is used to test the basic link, and the actual user patch cords are used to connect the tester to the channel or channel link. Essentially, the basic link is intended to provide more headroom than the channel, because it eliminates the uncontrolled variables of patch and transition cords. However, it still includes the more controlled variables of the 2-meter test cord connections at each end.
Many professional installers are moving to the use of a "permanent-link" test configuration. The permanent-link configuration essentially represents a logical connection that runs from the 8-pin modular connector on one end of the channel to the 8-pin modular connector on the channel's other end. To work with the permanent-link setup, the test equipment must have the ability to logically remove the variability of the test patch cords from the channel measurements, thereby eliminating the possibility of worn or damaged test cords that would adversely affect test results.
To date, the use of more expensive vector-measurement testers have typically been necessary to conduct permanent-link measurements. However, recent alternative methods have shown success by mating the tester as a matched set with a specific adapter cord, which is then factored into the tester's factory setup and calibration. These alternative methods are making the use of permanent-link measurements a more cost-effective and practical option for widespread use.
Higher-level network traffic
The professional cabling installer needs a full range of tools, from simple cable verifiers to multifeatured testers to LAN traffic generators/analyzers. Today, it is no longer sufficient to simply focus on the integrity of the network's physical layer, which provides the system's basic structured wiring links. As bandwidth specifications for various network protocols continue to escalate, performance headroom margins are steadily being squeezed for even the highest classifications of physical media. Regardless of whether the media includes twisted-pair, coaxial, or fiber-optic cabling, the ultimate test of its effectiveness resides in the network's ability to reliably transmit the required data under the stress of real-world operating conditions.
To assess a network's real-world performance capabilities, it is important to be able to selectively generate, capture, and analyze controlled traffic conditions for the OSI's second, third, and fourth layers-the data link, network, and transport layers, respectively. To manage and optimize ongoing network performance, it is also important to continually monitor and analyze actual traffic conditions across the entire network topology.
Testing and documenting a network's performance necessitates the ability to simulate a wide range of controlled traffic conditions, then capture and analyze the results. In addition, the ongoing evaluation of real-world network performance and troubleshooting of problems require the ability to monitor actual network traffic.
Working with real traffic
To test the full range of a network's capabilities, a user must be able to independently generate a broad spectrum of traffic conditions under tightly controlled conditions. Accomplishing that task means quickly saturating the network with large volumes of standard packet traffic and targeting various controlled traffic levels across a variety of bridged, switched, and routed environments. When it comes to testing the performance of a complex network environment, it is helpful for the traffic generator to be relatively portable, thereby enabling the operator to simulate the effects of introducing traffic spikes and peaks from various points and subnet segments within the overall topology.
After the simulated source traffic has been generated onto the network, the resultant destination traffic must be consistently captured, decoded, and logged before it can be analyzed. In a controlled performance-testing environment, a separate physical device connected in a remote location from the traffic generator is generally required to perform the traffic-capture function. Here again, when you are testing various aspects of a complex network, it is helpful to have a self-contained, portable traffic-capture device that can be easily connected at various points in the network topology.
In addition to the ability to simulate controlled conditions for assessing network-performance parameters, network managers also must possess the ability to monitor and analyze real-world traffic conditions, if they expect to effectively manage and optimize network performance on an ongoing basis.
Once the simulated or actual traffic has been captured, it must be analyzed at several levels so that network-performance levels can be accurately assessed. For self-contained, field-oriented testers, the analysis should focus primarily on critical performance parameters such as collision frequency, packet errors, and round-trip delay between specific points on the network. On the other hand, ongoing traffic-monitoring solutions must accumulate and analyze in-depth statistics on overall network-traffic levels, error rates, subnet loads, node-by-node usage patterns, and other traffic characteristics.
Answering the call
In response to their corporate customers' need for ensured performance levels, many professional cabling installers are beginning to offer some degree of performance testing as a value-added service. This service can significantly enhance the installer's competitive position. By selectively deploying portable, self-contained traffic-generator/analysis devices within their installation operations, these forward-looking contractors can stress the newly installed cabling to exceed actual LAN traffic levels. As a result, these contractors give their customers an added margin of confidence in the available performance headroom that the installation can provide.
For follow-up evaluations, upgrades, warranty service, and troubleshooting of installed networks, the ability to independently generate simulated traffic under controlled conditions can help to quickly distinguish between network-cabling problems and customer-equipment problems.
Network administrators also need easy-to-use, cost-effective solutions for the ongoing monitoring of actual traffic conditions. Instead of the field-oriented traffic-generator/analyzer capabilities that installers require, network administrators typically need more in-depth traffic monitoring and analysis solutions that can see across the entire spectrum of network activity under actual operating conditions.
In the final analysis, a reliance on industry-accepted standards for testing and certification, augmented by an analysis of the network's actual traffic-carrying capabilities, combine to form the foundation for ensuring a maintainable and extensible network.
G.W. Renken is senior RF engineer with Wavetek Wandel Goltermann (Research Triangle Park, NC).