When It Comes to Cable Test and Verification, Today's Technician Is Being Forced to Evolve, Investigate and Understand the Inner Workings of Network Protocols.
January 1, 2003
Here is a riddle: When are cable technicians not cable technicians? The answer is when they are forced to act as network test engineers. It’s not a clever riddle, but certainly indicative of the current state of affairs in new network installations.
Due to the increasing complexity of network systems, the increasing demand on complex protocols and services, and the severe physical requirements of new high-speed network connections the role of cable installation, test, and certification is evolving and becoming much more sophisticated.
Understanding these new physical and network-layer requirements of modern networks is the first step to surviving this shift in network implementations. The second is possessing the appropriate tools to examine the physical media according to these higher standards.
With the two together, you can provide simple and reliable network protocol and service verification to ensure that the network, once installed, will function as designed.
The physical network certification has always been important. Cable test and verification procedures have long been a part of the cable installation technician’s repertoire.
In fact, certification is the first, fundamental step to ensuring that a cable installation is acceptable, and when the network is handed over to the customer, everything works.
This certification, of course, occurs prior to the inevitable alterations by local IT departments and tech-savvy employees. Daily operations (and mishaps) within the wiring closet will inevitably affect the quality of the network.
In other words, test and certification procedures are also important to the customer who needs a working cable plant, and needs to know that they are building their complex enterprise systems atop a stable foundation.
They are equally important to the installation technicians themselves, helping to identify and replace trouble areas quickly so that installations run more smoothly. At the very least, proper certification documents the quality of their own work, providing a clean demarcation of responsibility when a new installation is handed over the customer.
However, networks have changed over the years, making the clean line a blurred one. While the design of a network cable infrastructure has certainly been simplified, as ring- and bus technologies gave way to more popular fiber-optic risers and point-to-point Ethernet connections.
Nonetheless, while the cable design may be simpler, the number of protocols, services, and devices that are critical to a network’s operation has grown. Domain Name servers, DHCP servers, and proxy servers are commonplace in enterprise networks.
Many networks possess Virtual Private Network capabilities (VPNs), intranets, extranets, and a host of security and encryption mechanisms and protocols to support it all. Simple Web servers have evolved into complex servers to support e-commerce, and may include entire clusters of servers, balanced across the network using yet more network sophistication.
While it is still necessary to certify the physical cable plant, the modern network demands additional scrutiny, if these types of services are going to work. The cable technician, as a result, is being forced to evolve, investigate and understand the inner workings of network protocols.
Consider the complex chain of events between an end-user, sitting in a cubicle, and a sales database used to generate customer quotes. At 8 a.m. the user logs in. This requires a connection to a DHCP server to get an appropriate IP address and other necessary configurations.
Next, a DNS lookup is required to locate the sales quote database — requiring communication between the end user’s PC and a DNS server. Once the connection can be made, checks need to be made to determine if the connection is allowed. This equates to separate communications to either proxy servers, authentication servers, Windows domain servers, VPN servers, or any combination of them all.
It is no longer good enough to verify that a cable will pass packets between an end-user and an upstream switch or router.
Physical-layer errors that may have been within acceptable limits in legacy systems, could create a bottleneck within the complex chain of events, and prevent a sales person from generating an order.
This means that the physical network needs to be tested against higher standards, and that the network must be examined as a whole; and yet it is inconceivable to test communication between each network drop and every available network service.
The solution requires increased training and expertise among cable technicians, which is time-consuming and costly or new tools, which can perform these extensive tests automatically.
These devices must not only certify cabling, but also possess the capabilities of protocol analyzers and be able to automatically detect and test these types of complex services. Where a handheld cable test unit would historically be used to verify that line interference and signal-to-noise ratios (SNR) were within acceptable limits (and perhaps even perform simple pings to verify network communication), these new tools will scan the network for critical devices and servers, verify their operation at the protocol layer, and even test performance to each of these key components.
The cables are tested and certified first to the much higher standards required for gigabit throughputs and the remaining tests are provided in simple pass/fail certifications. In other words, the inconceivable suite of tests discussed above becomes very conceivable.
The cable installation technician evolves into a network technician, and everyone benefits.
It’s important to understand, though, that the cable technicians are not — and should not — be replaced by higher-paid network service staff. While there is an undeniable need for cable technicians to expand their capabilities, the lowest common denominator of a new network’s success remains the same — the quality of the copper and fiber optic cabling.
However, the precision of the testing must be far greater, and the context of the testing has somewhat changed. Gigabit Ethernet over copper (1000Base-T) requires a much greater SNR margin than lower speed variations of Ethernet, as it is more susceptible to Near-end-Crosstalk (NEXT) and Far-end-Crosstalk (FEXT).
Explicitly, line noise is more likely to cause bit errors with these high-speed systems, and therefore the cabling infrastructure must be of a generally higher quality to present a cleaner signal.
Similarly, fiber optic cabling is susceptible to differential mode dispersion and signal attenuation that may effect bit error rates. However, due to the variances in fiber optics — i.e., LED versus laser light sources, core diameters, etc. — fiber optics present a challenge to certification tests under the narrow restrictions of gigabit systems.
Test equipment is required to first and foremost understand and account for these high standards, and enable the cable technician to determine without a doubt that the infrastructure is up to standard to support these extreme data rates. The additional complexity of what makes the network functional must be considered, tested, and certified.
At that time, specific lines should be tested to ensure that each of the network services mentioned earlier-DNS servers, DHCP, etc.-are responding appropriately.
As previously noted, there are many of these services; to avoid undue difficulty in verifying them all, a protocol analyzer should be used to automatically discover and test these devices, and to ensure there are no overlying IP configuration errors that may interfere with network operation.
By locating network services and servers, and then measuring their performance and reachability, higher-level problems can be identified at the outset, and will also greatly assist in network troubleshooting later on.
With the right approach, and the right tools, a complex, difficult and necessary task becomes much more manageable. As networks rely more and more on higher-level services, the initial certificat
ion of the network expands to include higher-level tests as well in order to compensate.
Another riddle: What do you get when you cross a cable tester with a network protocol analyzer? The answer to your prayers.
Again, not a particularly clever brainteaser, but true nonetheless. The protocol analyzer is a key tool in the successful diagnosis of a network; these devices are typically complicated and often very expensive.
A hybrid tool, that places the capabilities of data link and protocol testing into the hands of the cable technician, is more desirable. Cable test equipment remains as important-if not more, and can not be abandoned or replaced.
Networks have become so complex and precise that there is no longer room for bit errors. Networks are dependent upon new protocols and services that cannot accommodate the ill effects of a poor cabling plant.
A simple case of jitter or the occasional malformed packet can cascade through a complex sequence of network services and snowball into a major problem before the real network communication even begins.
Understanding these complex dependencies may be impossible to all but the highly trained (and expensive) network technicians. However, these physical and network-layer requirements can be easily examined using the appropriate test devices, placing network certification back into the hands of the cable tech.
The initial and most fundamental step in a new network implementation — the physical test and certification — has simply evolved; new technologies and equipment have been developed to remove the conundrum of convolution from the modern sophisticated network. After all, nobody likes riddles.
Chuck Ganimian is a product manager at Agilent Technologies Inc. for the FrameScope 350 product line. The following article first appeared in CompTIA’s Computing Channels magazine and has been reprinted with permission. For more information, visit www.comptia.org.