Wire and Cable may seem almost incidental, but they aren't. The nervous
system of your business merits close attention because it can have a profound
effect on success of your networking and communications efforts. This section
should give you basic information to discuss and negotiate your needs with the
cable guy. If you need greater detail,
Automation Access can create a detailed plan and specification for
The many types of cable formerly used in business buildings have boiled down to mainly just two: Category-5 (or better) UTP (Unshielded Twisted Pair) copper wire, and Optical Fiber. Devices that require coax cable connections, particularly IBM mainframe and midrange terminals, are now commonly interfaced to UTP wire through a device called a BALUN (Balanced to Unbalanced). Coax must still be used for some applications, particularly video cabling, but its topology (layout pattern) and reliability problems have pretty much banished it from the network.
The primary objectives of your wiring plan should be: reliability, adequacy,
and low cost for moves and additions. These objectives are best achieved by
adhering to a plan called "structured cabling".
The method by which this reliability and cost saving is accomplished is to pull only one type of wire which can be used for any variety of network, for telephone, and for serial terminals and printers. All configuration of the wiring system is done at a single point, a "wiring closet" on each floor to which every cable on that floor runs. Floor wiring closets are connected by a "riser cable".
Each workstation should have a minimum of two 4-pair Cat-5 cables: one for network and the other for telephone. These cables run from a wall jack at the workstation to a patch panel or backboard termination device (110 frame and blocks) in the wiring closet.
Four pair cable is practically universal. Even if your network uses only two pairs, and the cable contains four pairs, it is considered very bad form to run two network connections through that one cable. One reason is the possibility of interference between the two connections slowing your network, another is that, should you upgrade to a faster network that uses all four pairs. A third reason is that it can make things messy and difficult to interpret at the backboard or patch panel. Telephone lines can share the same cable without serious problems, except for upgrade and interpretability considerations.
In the wiring closet on each floor will be a "hub" to which each workstation network cable leads. Sometimes there is a server on each floor as well, which may or may not be in the wiring closet. Hubs, repeaters and other distribution devices are connected to each other through a "backbone cable" which may be Cat- 5, coax, or optical fiber depending on the unique needs of your installation.
The maximum run form any workstation to an active device (hub, repeater) is
100 meters (actual cable length).
The cost premium for plenum rated Cat-3 cable is very small, but due to a
worldwide shortage of plastic manufacturing capacity for Cat-5 plenum
insulation, the premium for Cat-5 plenum is about double non-plenum.
Problem was, he pulled two brands of Category-5 cable, and one brand was incompatible with the network cards in the clients computers. This was just a 10BaseT network with very short runs, but the workstations on that brand of cable simply didn't network. We had to bill the client for new network cards and the cost of figuring out why the old ones didn't work any more.
In a couple of other pre-pulled cases we found the wire so brittle we had trouble terminating it without breaking it. Again, this increased the client's expense. For these reasons we prefer to supply the cable, regardless of who is going to pull it.
If you identify a place where "we will never need a cable to here", that is the next place you will need a cable. Happens every time.
Ideally, both network cable and telephone cable should be pulled at the same
time, should be the same type of cable, and should run to the same backboard.
See Structured Cabling above.
In some areas where unions are strong, only a licensed electrician may install network cable. This is a problem, as most electrician have no idea whatever what data cabling is about and the job will often be totally botched. Most areas have more sensible legislators or weaker unions and anyone can pull "low voltage data cable" so long as they meet code.
Category-5 end-to-end requires close attention to bend radius, the twist of the wire pairs at the connections, and the exact hardware used. Cable pulls must be done neatly and with an eye to keeping cables away from sources of disrupting electro-magnetic interference, such as fluorescent fixtures. The wiring must be documented so it can be interpreted and reconfigured as needed.
Many clients presume they can get wire pulled cheaper than if we do it, so
we often let them use their own personnel, or their telephone guy, to pull the
cable (which we supply), then we do all the termination. This is still risky -
you'd be surprised how many wrong presumptions people can make about how the
job should be done. Cutting the cables off a foot too short can really mess
things up. (top)
TIA-568B specifies the color codes for the wire pairs and which pins on the jack the colors go to. Many devices have color guides printed on them for both 568A and 568B. Don't mix the two or you will have a nightmare installation.
At the backboard TIA-568B does not apply, because backboard termination treats the pairs in strict order of blue, orange, green and brown.
110 is a "punchdown" wire termination method. All the devices to which wire is terminated should be 110 punchdown devices because they can meet Category-5 specifications (if Cat-5 rated parts are used).
The familiar 66M punchdown blocks seen massed on telephone backboards are
replaced by a 110 frame, to which a termination block is fitted for each cable
terminated. We have a very successful 100BaseT network using the same
66M/RJ45 blocks we used for the 10BaseT network it replaced, but the wiring
between the punchdown tabs and RJ jacks does not meet the Cat-5 twist
requirement and could be a problem in a larger installation.
Even if fiber cost no more to install than copper, the cost of the network interfaces, hubs and other gear is much higher. Fiber is delicate and must have carefully controlled bends. The cost of splicing fiber is very much higher than for copper wire.
Primary advantages of optical fiber are transmission speed and distance. This has made it essential for many backbone installations connecting servers and hubs. With copper you are talking 100 meters, with fiber you can be talking thousands of meters. Speeds of over 100 megabits/second are still a bit dicey on copper, but are no problem for fiber.
Other advantages of fiber are security and electrical isolation. It is much harder to tap into a fiber transmission line than copper wire, which often doesn't even need to be touched to tap it. Since fiber does not conduct electricity, it is immune from grounding faults and other electrical problems that plague some networks.
For most uses, optical fiber must be pulled in pairs. One fiber is transmit, the other is receive.
A few years ago, FDDI (Fiber Distributed Data Interface) was the protocol that would take fiber to the desktop. Today a 100-Megabit/second FDDI network card for a PC costs about $1000 and a 100-Megabit/second Ethernet card costs about $24. Then ATM (Asynchronous Transfer Mode) was going to do it, but costs and Gigabit Ethernet put a stop to that.
The most common use of fiber in a business is to extend an Ethernet connection far beyond the distance it could go on copper wire. Wire to fiber conversion devices for this application are very common. The actual speed of transmission is no greater than over copper, so this is a distance play only.
ATM over fiber is used for high capacity campus backbones offering up to
2488-Megabit/second, but it is threatened by Gigabit Ethernet
(1000-Megabit/second) and 10-Gigabit Ethernet (not yet available). ATM over
fiber carries your DSL traffic from the DSLAM to the ISP's servers.
Look Ma! No wires! - Wireless Networking"Wireless" networking schemes abound, and every year is declared the "year of wireless" when wireless connections will take over. Problem: "Our wireless solution may cost more, but it's performance is not as good" is not the best selling point. You have to really need wireless to make it worthwhile.
Today, wireless networking is used primarily in warehouses, hospitals, and other places where people and equipment are highly mobile and must do data entry and retrieval while roving.
There's another problem with the futuristic image of the wireless office,
not yet significant only because wireless is not widely deployed. In every
office there are people very concerned about "radiation". What will they do
when they discover they are working in a sea of microwave transmitters?
Scientific evidence on electromagnetic radiation is still inconsistent and
inconclusive - they can make a case - and they will.
Outdoors is a major problem for infrared. Rain can greatly reduce the speed
of your network, and fog or very heavy rain can stop it entirely. Pigeons
landing on your transmission device can knock the alignment off.
Spread Spectrum is a technology originally invented for secure battlefield communications. It scatters transmission over a number of frequencies which can only be followed by a receiver tuned to the same code base. Low power spread spectrum wireless is used to "flood fill" warehouses, hospitals and other roving network applications. It does not require an FCC license to operate.
Microwave Transmission is most used to directly beam between
two locations. Its relatively high power requires FCC licensing.
©:Andrew Grygus - Automation Access -
Velocity Networks: Network Consulting Service - Internet Service Provider - Web Page Design and Hosting
All trademarks and trade names are recognized as property of their owners