An Emerging Old Technology


Power over Ethernet —

Wireless data is on the rise and the ability for copper wiring to efficiently conduct power and transmit data has kept copper wiring relevant — both in the inside plant and outside plant.

To help with standardization, the IEEE 802.3 Ethernet Working Group has issued two standards amendments since 2002; a third amendment, IEEE Std 802.3bt, is expected to be released in the 2nd half of 2018, adding to the breadth and capabilities of Power over Ethernet (PoE) technology.

Other trade organizations, such as the HDBASE-T alliance (, as well as individual manufacturers, have developed similar power-over-data technologies, some of them based on IEEE 802.3 standards, some of them foreshadowing the developing IEEE 802.3 standard, and some of them fully proprietary. As a result, many questions have arisen regarding this new technology. To help ensure safe installations and to add clarity to the connection requirements, the 2017 National Electric Code® introduced new regulations for wiring carrying both power and data, citing the growth of Power over Ethernet (PoE) technology.

This article provides background for powering over data networks, including the rise of PoE, discusses some of the major uses of PoE and how they fit into communications, reviews best practices, and finally, provides some insight to where PoE technology may be going.

Background of Power over Data Networks
While we might think of PoE as a new technology, its roots are as old as telephony itself. Alexander Graham Bell’s early work described the delivery of power and data simultaneously over a single wire in 1877. In US Patent 186,787 to Bell, he wrote: My invention has for its object, first, the transmission simultaneously of two or more musical notes or telegraphic signals along a single wire in either or in both directions, and with a single battery for the whole circuit, without the use of as many instruments as there are musical notes or telegraphic signals to be transmitted; second, the electrical transmission by the same means of articulate speech and sound of every kind, whether musical or not. (See Figure 1.)

Figure 1. Excerpt from US Patent 186,787 Bell, showing battery power delivered over telephony circuits.

Indeed, the history of power over data networks owes much of its lineage to telephony. Telephony introduced DC powering of the telephone circuit both of inline power and phantom powering, as well as the development of 4-wire circuits with phantom powering in connection with digital repeater-based systems. Powering technology reached into data communications as telephony evolved, both with line-powered ISDN and DSL technologies. At its root, for the phone company, the driving force for combining power and data over the same line remained the same: reliability and the ability to deliver voice and data over long distance. And for most of our lives the fact has been a given that the norm of lifeline communications services is independent of local power, and is enabled by line-powered systems commonly used in telephone outside plant engineering.

As technology moved forward to providing data networking in the 1990s, innovators in Ethernet networking technology began to apply Ethernet data networking technology to traditional telephony applications. These same innovators realized that lifeline powering would be expected of IP telephones, and that the technology developed for telephony could be adapted to provide both power and data over the new Ethernet networks. By early 1999, there were already several proprietary solutions in the market, both from data networking traditional telephony manufacturers. These systems utilized the basics of wireline powering technology.

Because the new PoE devices brought DC powering into the already-existing world of unpowered Ethernet data networks, each manufacturer came in with protocols to determine when and how much power to provide, including what limits to put on that power. To provide coherence to the various emerging powering techniques, in March 1999, the IEEE 802.3 Working Group began the initial steps towards its first PoE standard: IEEE 802.3af. The “call for interest” initiating the work of a study group on PoE (or “DTE Power via MDI” as it was known) already envisioned the usefulness of the powering technology beyond telephony to include wireless access points and video cameras (webcam) technology.

The IEEE 802.3af standard provided unified, interoperable specifications for detecting that a device required and could accept PoE, sensing how much power was needed, and limiting the voltage and current delivered. The IEEE 802.3af standard issued its first draft in mid-2001, and by the standard’s final approval in mid-2003, devices based on the IEEE PoE standard had been shipping for a while.

It took almost no time for the standard to be augmented to allow higher power and new management capabilities, and the work leading to the IEEE 802.3at standard began in late 2004 and was completed in 2009. As expected, market adoption increased, and the market demanded both more power and greater efficiency in delivery. As a result, the latest IEEE PoE standard, to be IEEE 802.3bt, began in 2013 and is expected to be released in the 2nd half of 2018. (See Figure 2.)

Figure 2. IEEE 802.3 Power over Ethernet Types and Classes (Courtesy Ethernet Alliance)

PoE’s Major Uses
Early PoE solutions, including those based on IEEE 802.3af, were limited to powering over 2 pairs of the 4 pairs commonly found in the balanced cables specified in TIA 568-C and used for Ethernet local area networks. IEEE 802.3af systems delivered up to 13 Watts to the powered device using DC levels of between 44 and 57 Volts, well within the safe electrically low voltage (SELV) levels. And with current levels of less than 175 mA per conductor, cable heating was not a significant concern.

These systems provided power for initial IP telephony applications, small webcams, access points, and other low power systems. PoE rapidly became a simple replacement for “wall wart” and other AC to DC transformers, delivering power and data to remote devices with a single cable.

IEEE 802.3at increased the power available to the remote device to 25.5 Watts both by tightening the minimum voltage sourced to 50 Volts (maintaining the maximum voltage at 57 Volts) and increasing the current levels to 300 mA per conductor, based on studies and inquiries involving the IEEE 802.3at Task Force, the cabling experts in TIA TR42, and consultations with Underwriters Laboratories (UL) for safety criteria. Powering at a level of 300mA per conductor was, and still is, considered the safe level even for large bundles of conventional 4-pair LAN cabling. Devices based on the IEEE 802.3at standard have extended this powering to all 4 pairs of copper, essentially doubling the power available at the same current level, in advance of the new IEEE 802.3bt standard. These standards set the stage for nearly any Ethernet-connected device to receive up to 51 Watts of power over the same cable used for data communications, often using existing, installed cabling.

While there is great diversity of PoE uses, including remote switches, access control, lighting, signs, speakers, and even such whimsical devices as PoE-enabled electric razors and guitars, the vast majority of PoE uses today remain in Ethernet network equipment, including IP telephones, wireless access points, and video cameras. These 3 types of equipment make up the lion’s share of PoE devices today, with more than 85% of the devices falling into one of those 3 categories in 2016, and approximately half of powered device shipments are IP telephones according to the Dell Oro Group.1 This includes both within-building, and, increasingly, outside the building extensions of local area networking technology. Still, the vast majority of legacy devices have communication as their primary function, whether it be for traditional voice, video monitoring, or relaying data connected to wireless devices.

PoE and Wireless
The growth of wireless technology has spurred the growth of wired Ethernet through PoE. Wireless access requires 2 primary connections to the wired network: one for powering, and one for the ultimate backhaul of the aggregated data into the network. Dreams of a wireless future envision wireless power in the last segment and wireless data backhaul; however, for now efficiency still favors wired power delivery and backhaul from fixed access points.

Electrical conductors remain the most efficient method for delivering power over a distance, and the work to install wiring pays off over years of operation. Using that installed wiring for data doubles the benefit, because, even though wireless has made great strides in data rate capacity and flexibility, as long as there is pressure on the last wireless network segment to provide faster data rates or support more users, reserving the wireless spectrum for the mobile segment and use fixed wired segments for backhaul is preferable. Because most infrastructure access points are fixed in location, wireless growth has synergistically driven the growth of the PoE-enabled wireless access point to the point where PoE access point shipments are nearly a third of PoE device shipments and growing.2

Moving forward, as Internet-driven connectivity adds functionality to more and more devices, everything from access control to lighting to industrial controls and sensors to kitchen appliances benefit from PoE functionality. The common unifying element is the need for data and power at a fixed location. With Internet-enabled functionality becoming widely pervasive, fixed devices are natural targets for Ethernet and PoE connection. In addition to network endpoints, sensors, and appliances, small network infrastructure elements such as low-port-count Ethernet switches are often PoE-enabled. A recent search by the Ethernet Alliance ( turned up PoE functionality on a wide variety of devices, including digital signage, health care consoles, LED lights, point-of-sale terminals, and even USB chargers. (See Figure 3.)

Figure 3. Evolution of Power over Ethernet (Courtesy Cisco Systems)

Best Practices
With PoE uses expanding so dramatically and power levels increasing, industry training on best practices for installing and managing PoE infrastructures is rising to meet the challenge. IEEE 802.3 specifications for PoE interfaces set this tone by providing detection and classification so the power sourcing equipment (PSE) can limit the power provided on circuits based on sensing the powered device. Best practices begin with ensuring that the devices are designed to the relevant standards for the intended application.

Figure 4. Ethernet Alliance PoE Certification Logo Mark for a Class 4 PSE (Courtesy of Ethernet Alliance)

IEEE 802.3 devices deliver power limited according to specifications designed to align with NEC® and other safety guidelines. IEEE 802.3 compliant power sources (PSEs) are also required to conform to UL specifications such as UL 60950-1 (this conformation is expected to include UL 62368 in the future). Best practices will include being able to identify these systems. To this end, the Ethernet Alliance, an industry consortium dedicated to the continued success and advancement of Ethernet technologies (, has developed a PoE Certification program based around the IEEE 802.3 PoE standards requirements. Systems conformant to the Ethernet Alliance’s PoE test plan will be marked with a logo indicating whether they are a power source (PSE) or a powered device (PD), and the maximum power classification (level) supported. This will enable easy recognition of interoperable systems designed to industry standards. (See Figure 4.)

Best practices for cabling for PoE begin with using the industry-standard cabling (which is specified at 24AWG or larger) in the horizontal cabling spans. These spans are often the problematic parts, behind walls or hidden in conduits. Use of 24AWG cabling also agrees with the 2017 NEC® specifications for bundling of power and data cables based on wire gauge and maximum current in Table 725.144. The 2020 NEC® will have further revisions to help support deployment of PoE systems.3

When situations get more complex due to higher power levels or narrower gauge cabling, best practices for installation and inspection of the cabling have recently been updated by both TIA and ISO/IEC. The TIA and ISO/IEC cabling standards bodies have completed technical reports and specifications for installing cabling for PoE. TIA TSB-184-A, “Guidelines for Supporting Power Delivery Over Balanced Twisted-Pair Cabling”, along with ISO/IEC TS 29125 Edition 2, provide a wealth of information for managing temperature rise in cabling systems even when all cables are saturated to their maximum current, including models and extensive tables for temperature rise vs. current vs. cable category (which correlates with minimum allowed wire gauge). These standards agree with the UL fact finding report that rated current levels of 300mA or less per conductor found in the vast majority of IEEE 802.3 standards-compliant (and standards-based) PoE equipment are of no concern when industry-standard 24 AWG or larger horizontal cabling are used.

Best practices for smaller wire gauge cabling or cordage are that these should only be used with PoE under limited conditions such as short connections to equipment. In general, narrower than 26AWG cabling, even as cordage, is not recommended for PoE because of the higher resistance per unit length and hence expected spot heating. There is limited data available on heating and performance with narrow gauge cable, and TIA is working on best practices for 28AWG cordage.

Where higher power levels are needed, best practices recommend minimizing bundling and using cabling in bundles of 24 or fewer, or, using cabling which heats less (such as 22 or 23 AWG cabling). The 2017 National Electric Code® echoes these 2 ways of complying for these higher currents: either complying with the bundling vs. AWG vs. current guidelines in Table 725.144, or using cabling rated for the maximum amperage expected, known as LP cable.

Finally, to aid in all this, automated infrastructure management (AIM) systems and intelligently coordinated power (ICP) sources (essentially managed PoE switches) are being augmented to automatically avoid situations which may result in overheating. These cabling and energy management systems incorporate knowledge of the cable bundling and how the individual cables relate to power delivered over them. As such, they can prevent overheating before it starts, even when the user reconfigures the power source or cabling system arrangement.

While AIM and ICP systems are not yet included in standards or the NEC® for best practices, an Amendment to ISO/IEC 18598 (AIM) to add information including bundle identifiers for PoE, to enable management of PoE by bundle, cable, and location, has begun the process of standardization. Also, Energy Management Systems, such as Cisco EnergyWise®, have been managing PoE power distribution for some time. As PoE evolves, you can expect best practices to include the increased use of integrated automated systems managing it.

The Future of PoE
The evolution of PoE is at a critical point. Applications are expanding at a rapid rate, beyond traditional communications systems to include infrastructure and IoT applications such as lighting and a variety of building automation sensors. In addition to these applications, the types of powering over data cables are expanding. While this article focuses primarily on IEEE 802.3 standard PoE and IEEE 802.3 standards-based solutions, the success of PoE has laid the ground for more diversity in line-powering.

The success of PoE has provided fertile ground for the emergence of simple power injectors, without the detection, classification, and limiting safeguards, of standards-compliant PoE. These devices are generally less complex than standards-based PoE, and do little more than directly couple a conventional power supply to the LAN wiring. These devices can be found in a variety of low voltage application such as surveillance cameras, and are often labeled PoE, even though they do not conform to IEEE 802.3 PoE standards. Telling whether these devices will take root as the market learns to recognize standards-based PoE devices as a best practice is early.

In addition to these proprietary powering schemes, new standards for data transmission, powering, and power over data lines, continue to emerge. The IEEE 802.3 Ethernet Working Group recently issued its first set of single pair Ethernet standards, originally targeted to fill needs for short-distance (15m) applications in automotive environments. The new standards include IEEE 802.3bu-2016, which defines a new power over data line (PoDL) technique operating on a single pair from 12, 24 or 48 Volts DC. The powering technique is generally not called PoE, but telling whether it will be confused with PoE since it also is power over an Ethernet link is too early.

PoDL is currently being extended for use with longer-reach industrial applications at lower data rates such as 10Mbps. These technologies are expected to find use in process control, outside plant, and building automation applications, and are in the early stages of standardization. They will use a new single-pair cabling system,
and practices for the use of power and data over PoDL in building or outside plant environments have yet to emerge.

PoE will continue to grow and find more and more applications as a high-speed data and power architecture. To support this growth, industry standards from various organizations will continue to evolve along with technology. PoE will be an important network architectural tool and work with others in a designer’s toolbox such as deep fiber deployments and new wireless technologies in providing broadband services to customers.

1. Dell Oro Group, IEEE PoE Webinar, April 2016,

2. Dell Oro Group, April 2016

3. The National Electrical Code® and NEC® are registered trademarks of the National Fire Protection Association (NFPA).

About the Authors
Ernie Gallo is a Director at Ericsson. He has more than 36 years of experience in electrical protection, outside plant requirements, and testing. Ernie currently serves as Chair or Vice Chair in several IEEE and ATIS committees. He is a representative on the National Electrical Code, Electrical Safety in the Workplace, and National Electrical Safety Code. For more information, please email or visit

Dr. George Zimmerman is an independent consultant, with more than 25 years of experience on wireline applications and their intersection with cabling infrastructure. He has been at the forefront of all BASE-T Ethernet faster than 1Gb/s, single-pair Ethernets, and Power over Ethernet. He is an active member of IEEE 802.3, TIA TR42, and NFPA. Currently, he is on the board of the NBASE-T Alliance, as well as Technical Committee Chair for the Ethernet Alliance. For more information, please email



About Author

Ernie Gallo is a Director -- Network Infrastructure Solutions, Ericsson. He has more than 38 years of experience in outside plant products, requirements, deployment, and electrical safety codes and standards such as the NEC and NESC. For more information, please email or visit Follow Ericsson on Twitter:

Comments are closed.