Fiber optics: a backbone for advanced building design
Fiber-optic cables are an integral part of a building communication system. Although they are commonly installed for the enterprise network communication, they are also designed into building-management systems and electrical-power coordination.
- Assess when and where to specify fiber-optic cable
- Classify the three different core sizes of the fiber-optic glass, which impact its performance.
- Summarize the cable listings recognized by the NEC and the scenarios of when the cable can be routed with power and other circuit types.
In the language of the information and communications technology (ICT) professional, fiber-optic cabling is the backbone medium for transporting data across campuses and through the spine of buildings. This in itself is not new. Fiber-optic cabling has played an integral part in network construction since the 1980s. Every year we continue to see advancements in technology, which in turn create more data and therefore highlight the role fiber optics plays for data transport. As a corollary to Moore’s Law, which describes the doubling of transistors per integrated circuit every 2 yr, there is a corresponding increase in the amount of data generated that then is transported.
In the 1980s, data transported from mainframe computers, mini-frame computers, and the early local area networks (LAN) connecting business workstations. In the 1990s, mainframes were rapidly being replaced by file servers communicating with desktop computers. E-mail and file transfer generated a lot of the network traffic. During this decade, access to the Internet and Web became available in the office. There was rapid growth in data transport as the content of the Internet evolved from text to graphics, audio, and occasional video. In the 2000s, video on the network continued to increase. We also saw the migration of telephone voice communications from private branch exchanges (PBXs) using large copper backbone telephone cables to Voice over Internet protocol (VoIP), which uses the data network infrastructure.
This brings us to the 2010s. Streaming video on the network is commonplace, and the Internet is so pervasive that it is not evident if data is being sourced from the enterprise network or the “cloud.” Access to office networks is expected to be ubiquitous in every building. Not having access can leave one feeling detached.
Through this evolution of computing and networking, there has been a parallel evolution in building design. It started with the automation of manufacturing and business processes; in other words, it started with the internal operations for which the buildings provided shelter. We are now witnessing the start of fully integrated and automated buildings. As we design advanced intelligent buildings, they are generating their own data to add to the network load.
The data generated by the building infrastructure is also related to the next emergence of data to be transported referred to as the Internet of Things. This term is attributed to Kevin Ashton, who in 1999 used it to describe embedded logic in physical items that can communicate with other devices beyond their own systems. Of course, the data generated by a building is only a fraction of what is expected to be generated by the occupants of the building. Everyday items occupants bring into the building will generate data. This is more than just tablet computers and smartphones. In the next few years we will see a new category of generated data from disposable logic devices. For example, imagine a disposable food container that has logic to identify when the contents is approaching its expiration date. The logic communicates with a smartphone that can upload the information to the cloud and then is accessed by an application on the smartphone.
Another technology component with large impacts on data is the proliferation of cameras in tablets and smartphones. This technology has a twofold impact. Video generates large data streams that are being used by more applications, and as the technology continues to improve for the devices, the video definition increases, which also increases the data stream size. All of this is generating more data, which is being backhauled (or transported) by the fiber-optic backbone.
Whether it is a route across town, between buildings on a campus, or between the communication rooms throughout a building, fiber-optic cable is the preferred backbone cable medium for data transport. Sending a signal is faster when using radio, but radio cannot match the bandwidth of optical fiber. Copper cabling is still an option for short data links, but cannot compare to fiber for the longer data channels. Fiber-optic cable is so widely used that for the purpose of this article, we will focus on cabling used for enterprise and facility building systems. This includes fiber-optic cable for LANs, passive optical networks (PON)s, control systems, monitoring, and security systems. To properly specify the cable, it is necessary to consider the type of communication link, the bandwidth requirements, strand count, and the environment in which the cable will be installed.
Optical fiber versus copper
Copper cable technology has worked hard to keep pace with the continuing increases in network bandwidth requirements. The current high-speed copper cabling standard in the U.S. is a Category 6A unshielded twisted-pair cable. It is rated at a frequency of 500 MHz and bandwidth that supports 10-GB/sec Ethernet for a standard cable length not exceeding about 100 m between network devices. This is more than adequate for most office desktop workstations and even backbone cabling in small buildings where there are one or two communications rooms not exceeding the 100-m cable length.
However, the predominant copper cabling being installed today is a Category 6 cable rated at 250-MHz frequency, which supports 1-GB/sec Ethernet. The bandwidth requirements at the desktop do not yet exist where we have passed the tipping point of requiring this higher bandwidth cable. For a typical hierarchal LAN design, copper cable to the office desktop is less expensive than a fiber-optic cable. Just as hard as the copper industry has worked to increase the bandwidth (i.e., longevity) of their product, the fiber optics industry is working to bring down the cost of their products.
There are a number of limitations for copper cabling, one being the 100-m channel distance. This ideally serves as the connection point between the communication-room network equipment and the desktop workstation (or another network-connected device). Copper cabling is more susceptible to electromagnetic interference (EMI) noise than fiber-optic cable, although our engineering team has routed unshielded Category 5e and 6 cable in high-EMI-generating manufacturing environments, and it has performed with no noticeable degradation in network performance.
Copper cable is susceptible to thermal noise generated in the cable from routing the cable in high-temperature environments. From a security standpoint, copper cable is less secure. The signals traveling along the cable generates radio frequency (RF) radiation which can be detected. However, fiber-optic cable is not impervious to signal tapping. Although a fiber-optic cable does not radiate an RF signal like a copper cable, a very skilled technician can expose a single strand of optical fiber in a cable, and the fiber can be bent to allow some of the light to escape. This may sound far-fetched, but fiber optic technicians that splice fiber-optic cable actually use a similar technique to inject light and siphon light from a fiber strand while splicing the cable to measure the quality of the splice.
Fiber-optic glass, which is an excellent dielectric, is effective in providing electrical isolation for the data circuits connected between different buildings on a campus or between communication rooms spread widely apart in a building. The entire cable can be made of dielectric materials, which requires no bonding or grounding. This can help in providing electrical noise isolation for the network equipment and to avoid ground loops.
Even though there is not a lot of Category 6A cabling being installed at the workstation level, within the Telecommunications Industry Association (TIA) Standards body there are people working on a Category 8 cable that is proposed to have four times the Ethernet bandwidth of the Category 6A cable. For now, and the foreseeable future, for a typical hierarchal LAN fiber-optic cable will continue to dominate the building backbone, and copper cable will continue to dominate the segment-to-the-workstation outlet. An exception to this is a non-hierarchal LAN called PON, or passive optical networking. In a PON network, the entire cable plant distribution consists of fiber-optic cable.
A true advantage copper cable has over fiber optics is the ability to transmit power from a communication room to a device. Power over Ethernet (PoE) is a technology allowing a variety of devices to be powered using the same data cable that is used for data transport. There is not enough power in a light signal in a fiber-optic cable to power a device. Fiber-optic end equipment still needs to rely on copper cables for this.
Communications and control
The largest use of fiber-optic cable in a typical commercial building is for enterprise networks. For industrial and manufacturing facilities, the networks supporting building and manufacturing automation may dominate the enterprise network. For other complex facilities such as a data center or an innovation hub, the building automation system (BAS) can be a significant portion of the enterprise network infrastructure, which is often completely isolated from the data center production network. Instead of a manufacturing-automation network, these types of facilities have networks supporting the data center’s data traffic and collaborative information, respectively.
For industrial and manufacturing facilities, the networks can be categorized into three major groups: office automation, facility automation, and factory automation. Office automation is what we are most familiar with; it includes such services as e-mail, time entry, billing, and other office utilities you would typically access from an office desk. The facility-automation networks provide data services for controlling and monitoring building equipment such as air-handling equipment, lighting control, power monitoring, UPS monitoring, and battery monitoring. For light industrial and commercial buildings, the facility-automation system would typically use a direct digital control (DDC) system.
In an industrial or manufacturing environment, complicated or custom requirements typically require the use of a programmable logic controller (PLC) or distributed control system (DCS). For DDC, PLC, and DCS data is generated for monitoring and control by a supervisory control and data-acquisition (SCADA) system.
The last network category is factory automation, which only exists for manufacturing environments. This includes data generated by manufacturing equipment directly or in “work in progress” stations associated with the manufacturing equipment. The data generated could relate to supervisory monitoring of the manufacturing tool, tracking inventory, tool metrology, or an alarm status.
What all of these networks have in common is that they are using Ethernet as the communication protocol and can be supported by the enterprise information technology (IT) network. The cabling infrastructure provided by IT can handle network protocols other than Ethernet, but the network equipment and architecture typically would not. The IT network may be able to provide a couple of fiber strands in the fiber backbone for a Modbus serial link, but this would not be data going through the IT Ethernet network equipment. IT staff members will design and support the network architecture to provide bandwidth allocation and network segregation between the different networks it hosts. They may choose to physically segregate the networks by allocating separate network equipment and fiber strands for each network type, or multiplex the data onto a few fiber strands using virtual LANs (VLANs) to isolate the network traffic. This is important to note because the network architecture is what determines the cabling architecture and, consequently, the type of fiber-optic cable along with the number of fiber strands that are needed.
Security systems such as access control and closed-circuit television (CCTV) may or may not be included in the enterprise networks. This depends on the building owner and the security plan. The trend in security equipment is to migrate away from analog and proprietary communication protocols to IP-based devices. Often the devices, such as cameras and card readers, are Power over Internet-enabled, allowing them to be powered by the same network cable that provides the data transport. Buildings fall at both ends of the spectrum—from completely segregated security networks, where they have their own network switches, rooms, and backbone fiber cable, to systems where security is just another device on the enterprise network with their own VLAN.
Facility-automation and factory-automation networks rarely reside on the enterprise network. In a control system, designing the SCADA may be on the enterprise network; but the peer-to-peer communications from PLC to PLC or PLC to remote input/output (I/O) panel is on a separate network often using industrial-rated network switches. Control networks may or may not use the Ethernet protocol for communications even though they are still connected with fiber-optic cable.