#
400ZR
An interoperable coherent optical interface designed to transport 400 Gb/s Ethernet over a single wavelength, primarily across point-to-point data center interconnect (DCI) links. Defined by the Optical Internetworking Forum (OIF), it is optimized for compact pluggable implementations, low power consumption, and transmission distances of approximately 100 km. Key specifications include a 60 GBaud symbol rate, 16QAM modulation, and concatenated forward error correction (CFEC). 400ZR signals occupy 75 GHz of spectrum.
800ZR
An interoperable coherent optical interface designed to transport 800 Gb/s Ethernet over a single wavelength for high-capacity DCI applications. 800ZR extends the principles of 400ZR to higher data rates using higher baud and enhanced FEC. Key specifications include a 118 GBaud symbol rate, 16QAM modulation, and open forward error correction (OFEC). 800ZR signals occupy 150 GHz of optical spectrum.
1600ZR
An interoperable coherent optical interface designed to transport 1.6 Tb/s Ethernet over a single wavelength for high-capacity DCI applications. 1600ZR extends the principles of 400ZR to higher data rates using higher baud. Key specifications include a 236 GBaud symbol rate, 16QAM modulation, and OFEC. 1600ZR signals occupy 300 GHz of optical spectrum.
1600ZR+
An interoperable coherent optical interface designed to transport 1.6 Tb/s Ethernet over a single wavelength for multi-span applications. 1600ZR+ extends the reach of 1600ZR from ~100 km to ~1,000 km through a higher symbol rate and enhanced modulation and digital signal processing (DSP) techniques. Key specifications include a 252 GBaud symbol rate, PCS-16QAM modulation, and OFEC. 1600ZR+ signals occupy 300 GHz of optical spectrum.
A
add/drop site
Location where optical wavelengths (data traffic) can be accessed (added to or dropped from the network).
agentic AI
Artificial intelligence capable of planning, making decisions, using tools, and taking actions with a degree of autonomy to achieve defined goals. Unlike traditional AI systems that primarily generate recommendations or responses, agentic AI can coordinate multiple steps and adapt its actions based on changing conditions and observed outcomes.
agentic operations
An operational approach in which AI agents analyze conditions, make decisions, coordinate actions, and execute workflows with limited human intervention. In network operations, this may enable faults, performance issues, and capacity requirements to be investigated and addressed across multiple systems through goal-driven automation.
AI
The capability of computer systems to learn from data, recognize patterns, reason, and make decisions that would traditionally require human intelligence. AI encompasses technologies such as machine learning, deep learning, generative AI, and agentic AI, and is increasingly used to improve network planning, operations, automation, and service assurance through data-driven analysis and decision-making.
AI accelerator
Specialized computing hardware designed to execute AI and machine-learning workloads more efficiently than general-purpose processors. AI accelerators, including GPUs and purpose-built processors (like TPUs), require high-bandwidth, low-latency networks to exchange large volumes of data during distributed training and inference.
AI cluster
A group of interconnected computing systems that work together to train or run AI models. AI cluster performance depends heavily on its networking fabric, which must efficiently move data among accelerators, storage systems, and other computing resources.
AI fabric
A high-performance network that interconnects processors, AI accelerators, storage, and other infrastructure supporting distributed AI workloads. AI fabric is typically designed to provide high bandwidth, low latency, predictable performance, and effective congestion management at very large scale.
AI infrastructure
The compute, networking, storage, software, power, and cooling resources required to develop, train, deploy, and operate AI systems. Optical connectivity is increasingly important within and between AI data centers because it supports the capacity, reach, and energy efficiency required by large AI workloads.
AIOps
The application of AI and machine learning to operational data to detect issues, identify causes, predict conditions, and automate network or IT operations. AIOps can correlate telemetry, alarms, topology, and historical data to provide faster and more actionable operational insight.
AI training
The process of using data and computational resources to adjust an AI model so it can recognize patterns and perform defined tasks. Large-scale training distributes computations across many accelerators, creating substantial requirements for low-latency communication and high-capacity data movement.
alarm indication signal (AIS)
A maintenance signal transmitted downstream to indicate that an upstream fault or defect has been detected. AIS notifies downstream equipment of the failure and suppresses secondary alarms that would otherwise result from the interruption of the original signal.
American National Standards Institute (ANSI)
The coordinating body for voluntary standards groups within the United States. ANSI is a member of the International Organization for Standardization (ISO).
amplifier
A device used to boost the optical signal to extend transmission distance. Also see optical amplifier.
amplitude modulation (AM)
A modulation technique that encodes information by varying the amplitude (intensity) of an optical signal.
analog bandwidth
The range of frequencies over which an analog device or channel can accurately transmit signals. In coherent optics, when describing modulators, drivers, TIAs, ADCs, and DACs, analog bandwidth determines how high a symbol rate (baud) the hardware can support, which directly influences the potential maximum data rate that can be supported. For example, a 100 GHz modulator enables a 200 GBaud coherent design.
application programming interface (API)
A defined set of rules and interfaces that allows software applications and systems to communicate and exchange data or invoke functions programmatically. In networking, APIs enable external applications, controllers, and orchestration systems to configure, monitor, and automate network devices and services.
attenuation
The reduction in optical signal power as it propagates through a transmission medium, such as an optical fiber. In optical networks, attenuation is typically expressed in decibels (dB) or decibels per kilometer (dB/km).
availability
The percentage of time that a system, service, or network is operational and accessible during the period that it is expected to be available. Availability is commonly expressed as a percentage over a specified period.
B
backbone
The core portion of a network that interconnects network segments or sites and carries traffic between them. A backbone typically provides high-capacity, high-performance connectivity across the network.
back-end network
A high-performance network that interconnects processors or AI accelerators within a computing cluster to support distributed training and inference workloads. It is typically optimized for very high throughput, low latency, minimal packet loss, and efficient collective communication among computing nodes.
bandwidth
Can refer to the range of frequencies that a transmission medium or device can support, expressed as the difference between its upper and lower frequency limits. Can also be a term used to indicate the amount of transmission or processing capacity, for example, network bandwidth.
baud or symbol rate
The symbol transmission rate of a communication signal, measured in symbols per second (Bd). It is not the same as bit rate, as each symbol can represent multiple bits depending on the modulation format.
bit error rate (BER)
The ratio of bits received in error to the total number of bits transmitted over a communication channel. BER is a key measure of transmission quality and is typically expressed as a ratio, such as 10⁻¹². Also referred to as bit error ratio. Pre-FEC BER indicates the raw error rate before forward error correction (FEC), while post-FEC BER indicates the remaining error rate after FEC.
bits per second (b/s)
The rate at which bits are transmitted or received each second. For example, an 800 Gb/s transceiver transmits 800 billion bits per second.
booster amplifier
See post-amplifier.
broadband
A high-capacity communications technology or service that supports the transmission of voice, video, data, and other digital traffic at high data rates. Historically, the term also referred to transmission over a wide range of frequencies that ;enables multiple signals to be carried simultaneously.
C
capacity optimization
The process of analyzing and adjusting network resources to maximize usable capacity while maintaining required performance and reliability. In optical networks, this may include changing modulation formats, adjusting channel rates, reorganizing spectrum, or identifying available margin for additional capacity.
carrier
An organization that owns or operates telecommunications infrastructure and provides network connectivity and communications services to residential or enterprise customers or other service providers.
central office (CO)
A service provider facility that houses network equipment used to aggregate, switch, route, and transport communications traffic. Historically associated with the public telephone network, modern central offices also support broadband, optical, IP, and mobile network services.
channel
A single optical communication path carried on a specific wavelength or frequency within a fiber. In wavelength division multiplexing (WDM) systems, each channel occupies a unique wavelength, allowing multiple channels to be transmitted simultaneously over a single fiber.
channel margin
The headroom between the measured or estimated signal quality of an optical channel and the minimum performance required for reliable operation. It indicates how much additional impairment the channel can tolerate and, in advanced coherent systems, can help determine whether its data rate can be increased.
chromatic dispersion (CD)
The spreading of an optical pulse because different wavelengths travel at different speeds through an optical fiber due to variations in the fiber's effective refractive index. CD is the most common form of dispersion in single-mode optical networks. Also see dispersion.
circuit
A physical or logical communication path that carries voice, data, or other traffic between two or more endpoints. A circuit may be dedicated or shared, depending on the network technology.
class of service (CoS)
A traffic classification mechanism that groups network traffic into different service classes and assigns priorities for forwarding and queuing. CoS provides differentiated treatment of traffic but does not, by itself, guarantee bandwidth, latency, or packet delivery.
client signal
A service or data stream that is transported over an optical network. The client signal is mapped onto an optical wavelength for transmission. Common client signals include Ethernet, Fibre Channel, and Optical Transport Network (OTN).
closed-loop automation
An automated process that observes network conditions, determines an appropriate action, implements that action, and verifies the resulting outcome. Feedback from the network is used to confirm success, make further adjustments, or reverse an action if the intended result is not achieved.
cloud on-ramp
A network access point or service that provides direct, secure connectivity between an enterprise or service provider network and a cloud environment. It improves performance and predictability by bypassing the public internet and using dedicated packet and optical connectivity to access cloud resources.
coarse wavelength division multiplexing (CWDM)
A WDM technology that uses widely spaced optical wavelengths, typically 20 nm apart across the 1270 nm–1610 nm wavelength range, to transmit multiple optical signals over a single fiber. Commonly used for cost-effective metro and access network applications.
Coherent-Lite
A simplified form of coherent optical transmission designed for short-reach, high-capacity applications. It retains many of the benefits of coherent technology, including higher loss tolerance and improved spectral efficiency, while reducing power consumption, cost, and digital signal processing (DSP) complexity.
coherent optics
An optical transmission technology that uses advanced modulation formats to encode information in the amplitude, phase, and polarization of light. At the receiver, the incoming signal is mixed with a local oscillator laser and processed using DSP to recover the transmitted data and compensate for optical impairments. Compared with traditional direct-detect optical transmission, coherent optics enables significantly greater fiber capacity, spectral efficiency, and transmission reach.
coherent pluggable
A highly integrated coherent optical transceiver that combines a DSP, photonic components, and control electronics within a standardized, multi-source agreement (MSA)–defined pluggable form factor (such as QSFP-DD or OSFP). It enables compact power- and space-efficient deployment of coherent optical interfaces in routers, switches, and optical transport systems. When deployed directly in routers or switches, coherent pluggables can simplify network architectures and reduce the need for dedicated transponder equipment.
coherent transceiver
A device that integrates the optical and electronic functions required to transmit and receive coherent optical signals over fiber. A coherent transceiver typically combines lasers, photonic components, a DSP, and control electronics to support high-capacity optical transport over metro, regional, and long-haul networks. Coherent transceivers are available in standardized MSA-defined pluggable form factors or are integrated into transponders in optical transport equipment.
colocation
The practice of housing network equipment owned by multiple service providers or organizations within the same physical facility to enable interconnection, improve operational efficiency, or share infrastructure.
colorless, directionless, and contentionless ROADM (CDC ROADM)
A ROADM architecture that combines three key capabilities: colorless, allowing any supported wavelength to use any add/drop port; directionless, allowing wavelengths to be routed to any line direction; and contentionless, allowing the same wavelength to be added or dropped simultaneously on multiple ports without blocking. Together, these capabilities simplify wavelength assignment, enable remote service provisioning, increase operational flexibility, and support highly automated, scalable, and resilient optical networks.
colorless ROADM
A ROADM architecture in which an add/drop port is not permanently assigned to a specific wavelength. Any supported wavelength can be assigned to the port, enabling flexible deployment of tunable transceivers across multiple technology generations while simplifying wavelength planning and network operations.
Common Language Equipment Identifier (CLEI)
A standardized identifier assigned to telecommunications equipment to uniquely identify a specific product, hardware variant, or component throughout its lifecycle.
Common Management Interface Specification (CMIS)
An industry-standard management interface for pluggable optical modules, defining how host systems configure, monitor, and manage module functions and operating parameters. Developed by the Optical Internetworking Forum (OIF), CMIS is widely used with high-speed optical modules, including QSFP-DD and OSFP.
congestion management
The detection, prevention, and control of network congestion to maintain efficient traffic flow and application performance. In AI and data center networks, congestion-management mechanisms help prevent packet loss, unfair bandwidth allocation, and delays that can reduce distributed-computing performance.
contentionless ROADM
A ROADM architecture that allows the same wavelength to be added or dropped simultaneously through multiple ports without wavelength contention. This improves flexibility for service restoration, network reconfiguration, and multi-degree network designs.
co-packaged optics (CPO)
An architecture that places optical engines within the same package as, or immediately adjacent to, a switching or processing chip. By shortening high-speed electrical connections, CPO reduces power consumption and signal loss while enabling higher interface speeds and greater data input/output (I/O) density.
D
dark fiber
Installed optical fiber that is not carrying an active optical signal and can be made available for another organization to illuminate using its own transmission equipment. Dark fiber users control the optical technology, capacity, and operational design placed over the fiber infrastructure.
data center interconnect (DCI)
A network that connects two or more geographically separated data centers to enable the exchange of data, applications, and services. DCI typically uses high-capacity coherent optical transmission to support workload mobility, data replication, business continuity, and, increasingly, distributed AI architectures such as scale-across networking.
data communications equipment (DCE)
Equipment that provides the interface between data terminal equipment (DTE) and a communications network. A DCE performs functions such as signal conversion, encoding, clocking, and line interfacing to enable data transmission over the network. Examples include optical transceivers and network termination equipment.
data rate
The amount of data transmitted or received over a communication channel per unit of time, typically expressed in bits per second (bit/s) or its multiples (kbit/s, Mbit/s, Gbit/s, or Tbit/s).
data terminal equipment (DTE)
Equipment that serves as the source or destination of user data in a communications system. A DTE originates data for transmission, receives transmitted data, or both. Examples include computers, servers, routers, and terminals.
decibel (dB)
A logarithmic unit used to express the ratio between two power or signal levels. In telecommunications and optical networking, decibels are commonly used to measure gain, loss, attenuation, and signal-to-noise ratios.
degrade
A reduction in the performance or quality of a network, service, signal, or system caused by one or more performance parameters falling outside predefined limits.
demultiplexing
The process of separating a multiplexed signal into its original individual signals or channels at the receiving end of a communication system.
dense wavelength division multiplexing (DWDM)
A WDM technology that uses closely spaced optical wavelengths, typically defined by the ITU-T wavelength grid, to transmit multiple optical signals over a single fiber. DWDM is widely used for high-capacity metro, regional, long-haul, and submarine optical networks.
digital signal processor (DSP)
A specialized processor that uses advanced mathematical algorithms to generate, receive, and optimize communication signals. In coherent optical systems, the DSP enables coherent transmission by encoding and recovering data using advanced modulation formats, performing forward error correction (FEC), compensating for optical impairments, and continuously adapting signal performance to maximize wavelength capacity, transmission reach, and spectral efficiency.
directionless ROADM
A ROADM architecture in which an add/drop port is not permanently associated with a specific line direction. Wavelengths added or dropped through the port can be routed toward any available line direction, simplifying network reconfiguration and improving restoration flexibility.
dispersion
The spreading of an optical pulse as it propagates through an optical fiber, causing pulse broadening that can limit transmission distance and data rate. In optical networking, dispersion typically refers to chromatic dispersion (CD) unless otherwise specified. Other important forms include polarization mode dispersion (PMD) in single-mode fiber and modal dispersion in multimode fiber.
dual polarization (DP)
An optical transmission technique that uses two orthogonal polarization states of light to carry independent data signals simultaneously on the same wavelength. By transmitting information on both polarizations, DP doubles the information-carrying capacity for a given modulation format and symbol rate. It is widely used in coherent optical transmission systems, including DP-QPSK and DP-16QAM.
E
edge cloud
A distributed cloud environment that places computing and storage resources closer to users, devices, or data sources. Metro and access optical networks provide the high-capacity, low-latency connectivity needed to link edge locations with centralized data centers and cloud resources.
effective signal-to-noise ratio (ESNR)
A measure of the effective signal quality at a coherent receiver that accounts for noise and other transmission impairments that degrade signal recovery. ESNR provides an overall indication of received signal quality and is commonly used to assess coherent transmission performance and margin.
electronic integrated circuit (EIC)
An integrated circuit that performs electronic functions such as signal processing, control, modulation drive, and signal amplification. In optical transceivers and photonic systems, it can be integrated or packaged closely with a photonic integrated circuit (PIC) to provide the electronic functions required to transmit, receive, and process optical signals.
element management system (EMS)
A software platform that manages, monitors, provisions, and maintains network elements within a communications network. It provides functions such as configuration management, fault management, performance monitoring, security management, and software maintenance for devices within a vendor or technology domain.
erbium-doped fiber amplifier (EDFA)
An optical amplifier that uses erbium-doped optical fiber as the gain medium and one or more pump lasers to amplify optical signals directly, eliminating the need for optical-electrical-optical (OEO) conversion. EDFAs provide broadband optical gain across the C-band or L-band, enabling multiple dense wavelength division multiplexing (DWDM) wavelengths to be amplified simultaneously over long transmission distances.
errored seconds (ES)
A one-second interval in which one or more transmission errors or defect events are detected. ES is a key performance monitoring metric used to measure the quality and reliability of optical transport and other telecommunications networks.
Ethernet
A family of packet-based networking technologies standardized by the IEEE 802.3 working group for communication over copper and optical media. Ethernet supports transmission rates from 10 Mb/s to hundreds of gigabits per second and is widely used in enterprise, data center, and service provider networks to transport IP and other higher-layer protocols.
F
fiber-optic cable
A transmission medium consisting of one or more glass or plastic optical fibers that carry information as pulses of light generated by lasers or other optical sources. Fiber-optic cables provide high bandwidth, low attenuation, and immunity to electromagnetic interference, making them the preferred medium for high-speed communications over short and long distances.
fiber-to-the-building (FTTB)
Fiber is deployed to a commercial or multi-dwelling building, with the final connection to individual users provided within the building using fiber, Ethernet, or other media.
fiber-to-the-cabinet (FTTCab)
Fiber is deployed to a street cabinet serving a neighborhood, with the final connection to customers typically provided over existing copper or other access media.
fiber-to-the-curb (FTTC)
Fiber is deployed to a distribution point near the customer’s premises, with the final connection typically provided over existing copper or coaxial cable.
fiber-to-the-exchange (FTTEx)
Fiber is deployed to the local exchange or central office, with the remaining connection to customers provided over the existing access network, typically copper.
fiber-to-the-home (FTTH)
Fiber is deployed directly to a residence, business, or other customer premises, providing an end-to-end optical connection capable of delivering high-speed broadband and other communications services.
fiber-to-the-node/neighborhood (FTTN)
Fiber is deployed to a neighborhood node serving multiple customers, with the final connection typically provided over existing copper infrastructure using technologies such as VDSL or G.fast.
fiber-to-the-x (FTTx)
A collective term for broadband access architectures that use optical fiber to connect a telecommunications network to locations closer to, or directly at, the end user. The suffix “x” denotes the point where the optical fiber terminates.
Fibre Channel
A high-speed networking technology and family of standards designed primarily for storage area networks (SANs). Fibre Channel provides reliable, low-latency transport of storage traffic between servers, storage systems, and switches, and supports point-to-point, arbitrated-loop, and switched-fabric topologies.
Fibre Connection (FICON)
IBM’s Fibre Channel–based channel architecture for connecting IBM mainframes to storage systems and other input/output (I/O) devices. FICON succeeded Enterprise Systems Connection (ESCON).
Flexible Ethernet (FlexE)
An Ethernet technology that supports flexible allocation, aggregation, and channelization of Ethernet capacity independently of individual physical interface rates. It enables multiple Ethernet services to share groups of physical interfaces and can carry rates that do not directly match the standard Ethernet port speed.
flexible grid (FlexGrid)
An optical spectrum-allocation method that supports variable channel widths and spacing rather than fixed wavelength slots. FlexGrid allows spectrum to be matched more efficiently to signals with different symbol rates and modulation formats, improving utilization of the available fiber spectrum.
Flexible Optical Transport Network (FlexO)
An ITU-T–defined Optical Transport Network (OTN) interface that bonds multiple transport lanes to carry client signals at rates beyond OTU4 (100 Gb/s). FlexO enables scalable, interoperable connections between high-capacity optical transport systems.
form factor
The physical and mechanical specification of a device or module, including its size, shape, connector arrangement, and host interface. In optical networking, form factors such as SFP, QSFP, QSFP-DD, and OSFP define how transceiver modules physically and electrically interface with network equipment but do not by themselves define the optical transmission technology or data rate.
forward error correction (FEC)
An error control technique that adds redundant information to transmitted data, enabling the receiver to detect and correct certain transmission errors without requiring retransmission. FEC is widely used in optical communications to improve transmission reach, increase capacity, and enhance link reliability.
full-spectrum transponder
A high-capacity optical transponder that combines multiple high-baud coherent signals within the transponder and presents a single line port to the photonic line system. By eliminating the need to externally multiplex individual wavelengths, it simplifies deployment and optical line design while enabling rapid deployment of high point-to-point capacity over dedicated fiber infrastructure.
G
generalized signal-to-noise ratio (GSNR)
A measure of optical transmission quality that combines the effects of amplified spontaneous emission (ASE) noise and fiber nonlinear interference into a single equivalent signal-to-noise ratio. GSNR is used to estimate the performance and achievable reach of coherent optical signals across a transmission path.
generative AI
AI that learns patterns from data and uses them to create new content, such as text, images, audio, video, or software code. Generative AI models often require substantial compute, storage, and network capacity for both training and inference.
Gigabit Ethernet (GbE)
A family of Ethernet technologies standardized by IEEE 802.3 that provides packet-based communication over copper and optical media at nominal data rates measured in gigabits per second, including 1GbE, 10GbE, 100GbE, 400GbE, 800GbE, and beyond. GbE remains widely deployed in enterprise and telecommunications networks.
gigabits per second (Gb/s; not Gbps)
A unit of data transmission rate equal to one billion (10⁹) bits per second, commonly used to express the speed or capacity of network and communications links.
graphics processing unit (GPU)
A highly parallel processor originally developed for graphics processing and now widely used to accelerate AI and other computationally intensive workloads. Large GPU deployments require high-performance networks to coordinate processing and rapidly exchange model parameters and data.
GPU as a service (GPUaaS)
A cloud service model that provides customers with on-demand or dedicated access to GPU computing resources without requiring them to own and operate the underlying infrastructure. GPUaaS is commonly offered by neocloud and other cloud providers for AI training, inference, and high-performance computing workloads.
GPU cluster
A group of interconnected computing systems containing multiple GPUs that operate together on distributed AI or high-performance computing workloads. A GPU cluster’s effective performance depends on the ability of the network to keep GPUs supplied with data and synchronized with one another.
H
hollow-core fiber (HCF)
An optical fiber that guides most of the transmitted light through a hollow, air-filled core rather than solid glass. Because light travels faster in air than in glass, HCF significantly reduces propagation latency while also reducing optical nonlinearities and other fiber-induced transmission impairments.
hot swap
The ability to remove and replace a hardware component while a system remains powered on and operational, without requiring a shutdown or reboot.
hybrid cloud
A computing environment that integrates private infrastructure with one or more public cloud services and enables applications or data to operate across them. High-capacity, secure connectivity is required to provide consistent application performance and efficient data movement between the different environments.
hybrid Raman amplification
An optical amplification approach that combines distributed Raman amplification within the transmission fiber with discrete optical amplification, typically using an erbium-doped fiber amplifier (EDFA). By combining the improved noise performance of Raman amplification with the gain of an EDFA, hybrid Raman amplification can improve optical signal-to-noise ratio (OSNR) and extend transmission reach in long-haul and high-capacity optical networks.
hyper-rail
See multi-rail.
hyperscaler
A cloud or digital service provider that operates large, globally distributed, and highly scalable computing, storage, and network infrastructure. Examples include Amazon Web Services (AWS), Microsoft, Google, Meta, and Alibaba, which are major users of high-capacity optical connectivity within data centers, between campuses, and across metro, regional, submarine, and long-haul networks.
I
in-band (transmission)
The transmission of management, control, or other protocols over the same communication channel as the user data, rather than over a separate management network.
inference
The process of applying a trained AI model to new input data to generate a prediction, decision, or response. Inference workloads may be distributed across centralized, regional, or edge data centers depending on their capacity, cost, and latency requirements.
inference cluster
A group of interconnected computing systems optimized to run trained AI models and serve inference requests at scale. Its network must support rapid access to models and data while meeting the response-time and availability requirements of deployed applications.
InfiniBand
A high-performance networking technology that provides low-latency, high-throughput communication and remote direct memory access between computing and storage systems. InfiniBand is widely used in AI and high-performance computing clusters where efficient accelerator-to-accelerator communication is critical.
information technology (IT)
The field encompassing the design, deployment, operation, and management of computer systems, software, networks, cloud services, data storage, and cybersecurity technologies used to create, process, store, transmit, and protect information.
infrastructure
The physical and virtual assets that support the operation of a communications network or other system. In networking, infrastructure includes fiber-optic cables, network equipment, data centers, power systems, software platforms, and related facilities.
in-line amplifier (ILA)
An optical amplifier installed between terminal sites to compensate for fiber loss and maintain optical signal power over long-distance links without terminating the carried services. An ILA typically amplifies multiple wavelength channels simultaneously and may also support monitoring, supervisory communications, and remote management. Also called an optical line amplifier (OLA).
intensity modulation with direct detection (IM-DD)
An optical transmission technique in which information is encoded by varying the intensity or power of an optical carrier and recovered at the receiver by directly detecting the received optical power with a photodetector. Unlike coherent transmission, IM-DD does not use a local oscillator to recover the phase and polarization of the optical signal. IM-DD is generally simpler and lower power than coherent transmission but provides lower spectral efficiency and transmission reach.
intent-based networking
A network-management approach that translates desired business or service outcomes into network policies, configurations, and automated actions. The network is continually monitored to determine whether the intended state is being maintained and to initiate corrective action when required.
intent-based operations
An operational approach that manages network operations based on specified outcomes rather than detailed procedural instructions. It combines intent, analytics, automation, and assurance to adapt network behavior while maintaining defined service and policy objectives.
interexchange carrier (IXC)
A telecommunications service provider that transports voice and data communications between local exchange areas or geographic regions over regional, national, or international backbone networks.
International Telecommunication Union (ITU)
A specialized agency of the United Nations responsible for developing international standards, allocating global radio spectrum and satellite orbits, and promoting the interoperability and development of telecommunications and information and communication technologies. The ITU publishes globally recognized standards through its ITU-T sector, including many used in optical transport and networking, such as the Optical Transport Network (OTN).
Internet Protocol version 6 (IPv6)
The current version of IP standardized by the Internet Engineering Task Force (IETF) for routing and addressing packets across IP networks. IPv6 uses 128-bit addresses, providing a vastly expanded address space compared with IPv4, along with improvements in routing, address configuration, and network scalability.
IP over dense wavelength division multiplexing (IPoDWDM)
An architecture in which coherent optical interfaces are integrated directly into IP routers or switches for transmission over a DWDM optical line system. It can simplify network layers and reduce equipment requirements, although operational responsibilities and optical performance must still be coordinated across the IP and photonic domains.
J
jitter
Short-term variations in the timing of a digital signal from its ideal or expected position in time. Excessive jitter can degrade signal quality and increase transmission errors.
K
L
large language model (LLM)
An AI model trained on large volumes of data to understand, generate, and reason with human language and other content forms. Training and operating large models can require thousands of interconnected accelerators and substantial data center and optical network capacity.
laser
A device that generates a highly directional, coherent, and nearly monochromatic beam of light through stimulated emission. Lasers are widely used as optical sources in fiber-optic communication systems.
latency
The time required for data or a signal to travel from a source to a destination. In optical networks, latency is primarily determined by the propagation delay of light through optical fiber, along with processing, switching, and routing delays introduced by network equipment.
leased line
A dedicated telecommunications connection provided by a service provider for the exclusive use of a customer. Unlike shared access services, a leased line provides dedicated, continuously available bandwidth between two or more locations.
line amplifier
See in-line amplifier.
line side
The side of a network element that interfaces with the transmission network, carrying signals toward or from other network elements over the optical line. In optical transport systems, the line side typically carries wavelength-based signals over the fiber transmission network, as opposed to the client side, which interfaces with client equipment such as switches or routers.
line terminating equipment (LTE)
Network equipment that originates and/or terminates line signals on the line side of a communications network, providing the interface between transmission facilities and client or switching equipment.
lit fiber
Optical fiber equipped with active transmission equipment and carrying optical communications signals. Unlike dark fiber, lit fiber services are generally delivered with the optical equipment, capacity, monitoring, and operational support managed by a service provider.
local area network (LAN)
A communications network that connects computers, devices, and other network resources within a limited geographic area, such as a home, office, building, or campus. LANs typically use Ethernet and/or Wi-Fi technologies to provide high-speed local connectivity and may connect to larger metropolitan area networks (MANs) or wide area networks (WANs).
lossless Ethernet
An Ethernet network designed to minimize or prevent packet loss through congestion control, traffic management, and flow-control mechanisms. Lossless Ethernet is commonly used for storage, remote direct memory access (RDMA), and AI workloads that are sensitive to packet loss and the delays associated with retransmission.
loss of frame (LOF)
A defect condition indicating that a network device has lost synchronization with the framing of an incoming digital signal. LOF typically results from signal degradation or upstream failures and prevents the proper reception of transmitted data.
loss of signal (LOS)
A defect condition indicating that a network device can no longer detect a valid incoming optical or electrical signal. LOS typically results from a fiber cut, disconnected cable, equipment failure, or severe signal degradation.
M
machine learning (ML)
A form of AI that enables computer systems to learn patterns from data and improve task performance without being explicitly programmed for every condition. In network operations, ML can support anomaly detection, traffic forecasting, performance optimization, and predictive maintenance.
managed optical fiber network (MOFN)
A dedicated optical network model commonly used by hyperscalers in which a service provider supplies and manages the fiber infrastructure and optical transmission equipment for the customer. A MOFN provides dedicated, scalable optical capacity with greater operational support than dark fiber while offering more customer control and flexibility than a conventional managed bandwidth service.
megabits per second (Mb/s; not Mbps)
A unit of data transmission rate equal to one million (10⁶) bits per second, commonly used to express the speed or capacity of network and communications links.
metro data center interconnect (metro DCI)
High-capacity connectivity between data centers located within the same city or metropolitan region, typically within 100 km. These links commonly prioritize compact equipment, low power, operational simplicity, and high capacity over the extreme reach required in long-haul networks.
metro network
A communications network that connects users, sites, data centers, and network facilities across a city or metropolitan region. Metro networks aggregate traffic and provide high-capacity connectivity between access locations, data centers, cloud on-ramps, and regional or long-haul networks.
modal dispersion
Pulse broadening in multimode fiber caused by different propagation modes arriving at different times because they travel along different paths. Modal dispersion is generally not present in single-mode fiber. Also see dispersion.
modulation format
The method used to encode information onto a carrier signal by varying properties such as amplitude, phase, frequency, or polarization. In coherent optical transmission, the selected modulation format affects data rate, spectral efficiency, optical reach, and tolerance to noise and other impairments.
multi-cloud
The use of cloud services from two or more cloud providers within a common application or IT environment. Network connectivity must provide secure, predictable access among users, enterprise locations, data centers, and the participating cloud platforms.
multimode fiber
An optical fiber with a larger core diameter than single-mode fiber that supports the propagation of multiple modes (light paths) of light. Multimode fiber is commonly used for short-distance, high-bandwidth communications within buildings, campuses, and data centers.
multiple system operator (MSO)
A service provider that operates multiple cable television and broadband networks, typically providing internet, video, voice, and other communications services to residential and business customers.
multiplex
To combine multiple optical signals, wavelengths, or client data streams for transmission over a single optical fiber or communications channel.
multiplexer (MUX)
A network device that combines multiple client signals, data streams, or optical wavelengths into a single higher-capacity signal for transmission over a shared communications link or optical fiber.
multiplexing
The process of combining multiple client signals, data streams, or optical wavelengths for simultaneous transmission over a single communications link or optical fiber. Common multiplexing techniques include time-division multiplexing (TDM), wavelength-division multiplexing (WDM), and Optical Transport Network (OTN) multiplexing.
multi-rail
A high-density photonic networking architecture that integrates multiple optical rails (fiber pairs) within a single optical module. By carrying several independent optical paths over a single, highly integrated interface, multi-rail photonic systems greatly increase fiber density while simplifying deployment and enabling optical connectivity to scale efficiently for large AI computing infrastructures.
N
near-package optics (NPO)
An architecture that places optical engines close to, but outside, the package containing a switching or processing application-specific integrated circuit (ASIC). By shortening the high-speed electrical connections between the ASIC and the optics compared with traditional front-panel pluggable optics, NPO reduces electrical signal loss and power consumption while providing greater separation and serviceability than co-packaged optics.
neocloud
A cloud service provider focused primarily on accelerated computing infrastructure for AI and high-performance computing workloads, often delivered through GPU as a service and related services. Neocloud providers commonly deploy large GPU clusters and require high-capacity networks within, between, and beyond their data centers.
neoscaler
A provider that builds and operates large-scale, GPU-based AI infrastructure to deliver hyperscale-class compute, networking, and storage resources for AI workloads. Neoscalers focus primarily on AI training and inference rather than the broad portfolio of cloud services offered by traditional hyperscalers. Examples include CoreWeave, Nebius, Crusoe, Lambda, and Together AI.
network
A collection of interconnected nodes that exchange information over optical or electrical transmission media using communications equipment and standardized protocols. Networks support the transport of voice, data, video, and other digital services across local, metropolitan, and wide geographic areas.
Network Configuration Protocol (NETCONF)
A network management protocol standardized by the Internet Engineering Task Force (IETF) that enables applications and management systems to install, modify, retrieve, and delete the configuration of network devices. NETCONF commonly uses YANG data models to provide structured, programmable network management.
network digital twin
A dynamic digital representation of a physical network that uses topology, configuration, telemetry, and behavioral models to reproduce network conditions. A network digital twin can be used to simulate changes, predict outcomes, validate automation, and assess potential effects before actions are applied to the production network.
network element (NE)
An individually addressable device within a communications network that performs one or more transport, switching, routing, or management functions. Network elements can be monitored, configured, and managed through a network management system.
network management system (NMS)
A software platform that monitors, configures, and manages network devices and services. An NMS collects operational data, alarms, performance metrics, and inventory information to support network administration, optimization, and troubleshooting.
O
observability
The ability to understand the internal condition and behavior of a network or system by collecting and correlating telemetry, events, logs, topology, and other operational data. It goes beyond basic monitoring by helping operators investigate why a condition occurred and understand its effect on services.
octal small form-factor pluggable (OSFP)
A high-density, hot-pluggable transceiver form factor that supports eight electrical data lanes. Its larger physical size and greater thermal capacity accommodate higher-power modules and high-capacity interfaces such as 400G, 800G, and 1.6T, including coherent optical transmission technologies.
OpenROADM (open reconfigurable optical add/drop multiplexer)
An industry initiative that develops open, interoperable specifications for optical networking, enabling equipment from different vendors to operate within a common network. OpenROADM defines interfaces, data models, and parameters for optical network elements such as ROADMs, transponders, pluggable optics, and amplifiers, as well as models supporting integration with packet and routing equipment. It enables open, disaggregated optical networks and coordination between the optical and packet layers.
OpenROADM 800G
An interoperable coherent optical transmission mode defined by the OpenROADM MSA for carrying 800 Gb/s over metro and regional dense wavelength division multiplexing (DWDM) networks. It uses standardized probabilistic constellation shaping (PCS) and forward error correction (FEC) to provide greater reach than 800ZR while enabling interoperability between coherent transceivers from different vendors.
Open Systems Interconnection (OSI)
A seven-layer reference model developed by the International Organization for Standardization (ISO) that standardizes network communications functions and provides a common framework for understanding, designing, and troubleshooting networks. The seven layers are Physical (Layer 1), Data Link (Layer 2), Network (Layer 3), Transport (Layer 4), Session (Layer 5), Presentation (Layer 6), and Application (Layer 7). Optical transmission primarily operates Layer 1, while IP routing operates at Layer 3.
OpenZR+
An interoperable coherent pluggable specification that extends ZR-based technology with stronger FEC to support longer or multi-span 400G optical links. It is intended to combine router-compatible pluggable form factors with greater optical performance than basic 400ZR implementations.
operational intelligence
Actionable insight produced by analyzing real-time and historical operational data to support decisions, optimization, and automation. In optical networks, operational intelligence can transform telemetry and performance measurements into an understanding of channel health, available margin, emerging risks, and potential corrective actions.
operations support system (OSS)
A software platform or suite of applications that supports the operational lifecycle of communications networks and services. OSS functions typically include service provisioning, network inventory, fault management, performance monitoring, service assurance, and automation.
optical add/drop multiplexer (OADM)
A network element that allows selected wavelengths to be added to or dropped from a multi-wavelength (WDM) optical signal while allowing the remaining wavelengths to pass through without being terminated.
optical amplifier (OA)
A device used to directly amplify an optical signal without converting it to an electrical signal. OAs can be deployed as post-amplifiers (booster amplifiers), in-line amplifiers, or pre-amplifiers, depending on their location and function within an optical link.
optical channel
An optical signal transmitted over a specific wavelength or frequency slot within an optical fiber. In a wavelength division multiplexing (WDM) system, each optical channel carries data independently of other channels and can be individually added, dropped, switched, or routed through the optical network.
optical circuit switch (OCS)
A device that establishes reconfigurable optical connections between input and output fibers or ports without converting the carried signals into electrical form. An OCS dynamically switches optical paths while remaining transparent to the wavelength, modulation format, data rate, and protocol of the optical signals.
optical control plane
The set of protocols and software that automatically establishes, maintains, optimizes, and restores optical network connections. The optical control plane enables functions including path computation, signaling, resource discovery, and protection or restoration following network failures.
optical data unit (ODU)
An Optical Transport Network (OTN) information structure that carries client data and path-related overhead through an OTN. Different ODU types provide standardized containers for transporting, multiplexing, monitoring, and switching services at a range of data rates. For example, ODU4 operates at approximately 104.8 Gb/s and is commonly used to carry 100 Gb/s client signals.
optical data unit-Cn (ODUCn)
A flexible-rate OTN data unit used to carry client information within an OTUCn signal, where n represents the number of approximately 100 Gb/s capacity increments. For example, ODUC4 provides approximately 400 Gb/s of OTN payload capacity. ODUCn supports high-capacity transport while retaining OTN capabilities such as multiplexing, monitoring, and end-to-end path management.
optical data unit flexible (ODUflex)
A variable-rate OTN data unit that provides transport capacity matched to the rate of a client signal. ODUflex allows OTN to efficiently carry services whose rates do not correspond to traditional fixed-rate ODU containers.
Optical Internetworking Forum (OIF)
An industry organization that develops interoperable implementation agreements and technical specifications for optical networking, including coherent optical interfaces, electrical interfaces, and related technologies. OIF specifications are widely used to promote interoperability among equipment and component vendors, including standards such as 400ZR, 800ZR, 1600ZR, and 1600ZR+.
optical line amplifier (OLA)
See in-line amplifier.
optical margin
The performance headroom of an optical connection above the minimum required for reliable operation, representing its tolerance for additional impairments or degradation due to factors such as noise, loss, aging, fiber repairs, and component variations.
optical path
The route over which an optical signal or optical channel is transmitted between endpoints in an optical network. An optical path may traverse multiple fiber spans and photonic network elements, such as amplifiers, ROADMs, and optical switches.
optical service channel (OSC)
A dedicated optical communications channel used to transport network management, control, monitoring, and maintenance information between optical network elements. The OSC is transmitted over the same fiber as the traffic-carrying optical channels but operates on a separate wavelength.
optical signal-to-noise ratio (OSNR)
A measure of optical signal quality defined as the ratio of optical signal power to optical noise power measured in a 0.1 nm reference bandwidth (approximately 12.5 GHz), typically expressed in decibels (dB). In DWDM and coherent optical networks, OSNR is a key performance metric used to determine whether an optical channel can support its intended modulation format, data rate, and transmission reach.
optical span
A section of an optical link between two adjacent optical amplification or termination points, such as amplifiers, ROADMs, or terminal sites, typically consisting of the connecting fiber and passive components. An optical span is characterized by properties including length, loss, dispersion, and optical return loss.
Optical Transport Network (OTN)
A standardized digital transport architecture defined by the ITU-T that provides the encapsulation, multiplexing, switching, management, and transport of client signals over optical networks. OTN supports a wide range of client services, including Ethernet, Fibre Channel, SONET/SDH, and other transport protocols.
optical transport unit (OTU)
An OTN information structure that adds framing, section-monitoring overhead, and FEC to an ODU for transmission between OTN interfaces. The OTU forms the line signal used across an OTN link or between regeneration points.
optical transport unit-Cn (OTUCn)
A flexible-rate OTN transport signal composed of n approximately 100 Gb/s capacity increments. It provides framing and transport functions for carrying an ODUCn signal over high-capacity interfaces. For example, OTUC4 provides an approximately 400 Gb/s OTN transport signal for carrying an ODUC4.
out-of-band (OOB; transmission)
The transmission of network management, control, or maintenance information over a communications channel that is separate from the client data traffic. OOB communication is commonly provided through a dedicated management network, Ethernet connection, or optical supervisory channel (OSC).
outside plant (OSP)
The physical communications infrastructure located outside buildings that supports the transmission of voice, data, and video services. The OSP includes fiber-optic and copper cables, ducts, poles, conduits, cabinets, splice closures, handholes, and other network facilities connecting central offices, data centers, cell sites, and customer premises. Active equipment deployed in OSP environments typically requires extended-temperature and environmentally hardened designs to operate reliably under varying outdoor conditions.
P
payload
The portion of a communications frame, packet, cell, or transport container that includes the user data or client information being transported. It excludes the protocol headers, trailers, and other overhead used to route, manage, and transmit the data.
performance monitoring (PM)
The continuous measurement and analysis of network or service performance to assess operational health, verify service quality, and detect degradation before it impacts service.
photonic integrated circuit (PIC)
A device that integrates multiple optical functions, such as light generation, modulation, switching, filtering, or detection, onto a single substrate. Integration can reduce the size, power consumption, cost, and complexity of optical communications equipment.
Poincaré sphere
A graphical representation of the state of polarization (SOP) of an optical signal. Each point on the sphere represents a unique polarization state, with linear polarization states along the equator, circular polarization states at the poles, and elliptical polarization states elsewhere. The Poincaré sphere is widely used to visualize and analyze polarization behavior in optical communication systems.
point of presence (PoP)
A physical location where a network service provider houses communications equipment to provide network connectivity, exchange traffic, or deliver services. A PoP typically contains routing, switching, optical transport, and other networking equipment that connects customers, partners, or other networks.
polarization mode dispersion (PMD)
Pulse broadening in single-mode fiber caused by slight differences in the propagation speed of two orthogonal polarization modes, typically resulting from fiber imperfections or mechanical stress. Also see dispersion.
port
A physical or logical connection point through which a device sends or receives data. A physical port provides an interface for connecting network equipment or transmission media, such as an Ethernet or Fibre Channel port. A logical port, such as a TCP or UDP port, identifies a specific application or network service.
post-amplifier (post-amp)
An optical amplifier placed near the transmitting end of an optical link to increase optical launch power and extend transmission reach. Also called a booster amplifier.
pre-amplifier
An optical amplifier placed near the receiving end of an optical link to increase the power of incoming optical signals and improve receiver sensitivity.
predictive maintenance
The use of operational data and analytics to identify equipment degradation and schedule maintenance before a failure occurs. In optical networks, trends in power, temperature, errors, component behavior, and signal quality can indicate when intervention is likely to be required.
predictive operations
The use of analytics, AI, and operational data to anticipate network conditions, service degradation, or failures. With predictive operations, operators or automated systems can act before customers are affected rather than responding only after alarms or outages occur.
probabilistic constellation shaping (PCS)
A modulation technique that transmits some constellation symbols more frequently than others to improve transmission efficiency. In coherent optical systems, PCS enables finer optimization of data rate, optical reach, and transmission performance based on the available optical channel conditions.
provider
An organization that delivers communications, networking, cloud, or infrastructure services to customers. Examples include telecommunications service providers, internet service providers, cloud service providers, and multiple system operators.
public network
A communications network operated by a service provider to deliver connectivity and communications services to multiple customers or the public. Public networks support services such as voice, data, internet access, and broadband connectivity.
Q
quadrature amplitude modulation (QAM)
A digital modulation technique that encodes data by varying the amplitude and phase of two orthogonal components of a carrier signal, known as the in-phase (I) and quadrature (Q) components. The resulting combinations of amplitude and phase form a constellation of symbols, enabling high data rates and efficient use of optical spectrum. For example, 16QAM uses 16 constellation points, allowing each symbol to represent four bits of information.
quadrature phase-shift keying (QPSK)
A digital modulation format that uses four distinct carrier-phase states to encode data. The four phase states allow each symbol to represent two bits of information. QPSK provides high tolerance to noise and transmission impairments, enabling longer optical reach than higher-order QAM formats, although it carries fewer bits per symbol.
quad small form-factor pluggable (QSFP)
A compact, hot-pluggable transceiver form factor that supports four electrical data lanes, enabling higher aggregate data rates than SFP. These modules can support various optical and electrical interfaces and transmission technologies.
quad small form-factor pluggable double-density (QSFP-DD)
A high-density, hot-pluggable transceiver form factor that supports eight electrical data lanes, doubling the lane density of QSFP. QSFP-DD is widely used for high-capacity interfaces such as 400G and 800G and can support both client and coherent optical transmission technologies.
quality of service (QoS)
A set of network capabilities used to prioritize and manage traffic to meet defined performance objectives, such as bandwidth, latency, jitter, and packet loss, for different applications and services.
R
radio frequency (RF)
A range of electromagnetic frequencies commonly associated with radio-wave transmission but also used for high-speed analog electrical signals in coherent optical systems. In coherent transceivers, RF signals are used to drive optical modulators and carry signals from coherent receivers, providing the high-speed electrical interface between the optical front end and signal-processing electronics.
rail
A fiber pair that provides one of several parallel optical transport paths in a multi-rail photonic network architecture. Each rail provides an independent optical path and a portion of the overall network capacity.
Raman amplification
An optical amplification technique that uses high-power pump light to create gain within the transmission fiber through stimulated Raman scattering. Because amplification is distributed along the fiber, Raman amplification can improve optical signal-to-noise ratio (OSNR) and extend transmission reach.
reconfigurable optical add/drop multiplexer (ROADM)
An optical networking device that dynamically routes, adds, or drops wavelength channels while keeping traffic in the optical domain. A ROADM can access and switch wavelengths across the optical spectrum on each line port, enabling flexible wavelength routing without optical-electrical-optical (OEO) conversion. By switching traffic entirely in the optical domain, ROADMs reduce power consumption and cost while enabling efficient fiber capacity use.
regeneration
The process of restoring a degraded optical signal by receiving and reconstructing the transmitted data before retransmitting it as a new optical signal. In optical transport networks, regeneration is typically performed through OEO conversion, providing re-amplification, reshaping, and retiming (3R) to restore signal quality and extend transmission reach.
repeater
An optical device used to extend transmission distance by periodically amplifying optical signals along a fiber link. Repeaters are widely used in submarine cable systems, where they are installed at intervals along the undersea cable to compensate for fiber loss and enable transmission over very long distances.
RESTCONF
A network management protocol standardized by the Internet Engineering Task Force (IETF) that provides a RESTful interface for accessing and modifying configuration and operational data defined by YANG data models. RESTCONF uses HTTP-based methods to enable applications and management systems to programmatically configure, monitor, and manage network devices and services.
restoration
The process of recovering network services after a failure by rerouting traffic over available network resources or alternate paths.
S
scale-across network
A high-capacity network that interconnects geographically separated AI data center campuses or regions to enable the training of a single AI model across multiple sites. Scale-across networks use networking and optical technologies appropriate to the required distance and capacity, with coherent optical transport typically used as distances extend to tens of kilometers and beyond.
scale-out network
A high-performance network that interconnects multiple AI compute systems or clusters within a data center or campus, enabling AI workloads to be distributed across a larger pool of accelerators. Scale-out networks typically use high-speed packet networking, such as Ethernet, to provide connectivity beyond the limits of a single scale-up domain.
scale-up network
A high-performance, ultra-low-latency network that interconnects accelerators, such as GPUs, within an AI compute system or tightly coupled compute domain. Scale-up networks enable accelerators to exchange data at high speeds and efficiently operate together on computationally intensive AI workloads.
service assurance
The processes and capabilities used to monitor, analyze, and maintain the availability, performance, and quality of network services. Service assurance connects network conditions with customer-facing services so that degradation can be detected, isolated, and addressed before or after service objectives are affected.
service-level agreement (SLA)
A formal agreement that defines the expected performance, availability, support, and other measurable commitments for a service. An SLA may specify metrics such as uptime, latency, repair time, packet loss, and consequences when agreed service levels are not met.
signaling
The exchange of control information between network devices to establish, modify, manage, and terminate network connections or services.
signal-to-noise ratio (SNR)
A measure of signal quality that compares the power of a desired signal to the power of noise, typically expressed in decibels (dB). SNR is a general performance metric used across many communications technologies to indicate how reliably information can be transmitted and received.
silicon photonics (SiPh)
A technology that uses silicon-based semiconductor manufacturing processes to integrate optical components and functions onto photonic integrated circuits (PICs). SiPh enables high-volume manufacturing and close integration of optical and electronic functions for transceivers, switches, and other communications equipment.
single-mode fiber (SMF)
An optical fiber with a solid glass, typically silica, core that is designed to support only the fundamental spatial mode of light over its intended operating wavelength range. By minimizing modal dispersion, SMF enables high-capacity, long-distance optical transmission.
small form-factor pluggable (SFP)
A compact hot-pluggable transceiver form factor that provides interchangeable optical or electrical interfaces for communications over fiber-optic or copper media. SFP modules can be installed or replaced without powering down the host device. Also see octal small form-factor pluggable, quad small form-factor pluggable, and quad small form-factor pluggable double-density.
software-defined networking (SDN)
A networking approach that separates network control functions from the underlying data-forwarding infrastructure and enables network behavior to be controlled programmatically. SDN supports centralized or logically centralized control, automation, and dynamic configuration of network resources and services.
spectrum defragmentation
The process of reorganizing optical channels to consolidate unused spectrum and create larger contiguous spectrum blocks. Spectrum defragmentation improves spectrum utilization and can make room for new high-baud or high-capacity services without requiring additional fiber.
state of polarization (SOP)
The orientation and form of the electric field of an optical signal as it propagates through a transmission medium. The SOP can change over time because of environmental and mechanical effects on the fiber and must be tracked and compensated for by coherent receivers.
storage area network (SAN)
A dedicated high-speed network that provides computer systems with access to shared storage resources. SANs typically provide block-level storage access and commonly use Fibre Channel, although other protocols and networking technologies can also be used. By separating storage traffic from general-purpose network traffic, they enable scalable, reliable, and high-performance access to centralized storage.
streaming telemetry
The automated, real-time or near real-time delivery of structured operational data from network devices to external monitoring or analytics systems. Unlike periodic polling, streaming telemetry pushes data as it becomes available or at configured intervals, providing timely, high-frequency updates that support network observability, anomaly detection, and closed-loop automation.
T
telecommunications management network (TMN)
An ITU-T framework that defines a standardized architecture and interfaces for monitoring, configuring, managing, and maintaining telecommunications networks and services.
telemetry
Operational data collected from network devices or systems and transmitted to monitoring, management, or analytics platforms to provide visibility into network state and performance. In optical networks, telemetry can include power levels, signal quality, errors, temperature, configuration state, spectrum usage, and other performance measurements.
terabits per second (Tb/s)
A unit of data transmission rate equal to one trillion (10¹²) bits per second. Commonly used to express the capacity or throughput of high-speed optical communications systems and networks.
throughput
The rate at which data is successfully transmitted over a communications network or system, typically measured in bits per second (b/s). Throughput represents the actual achieved data transfer rate and may be lower than the nominal capacity or line rate due to protocol overhead, network conditions, and other factors.
transceiver
A device that combines transmitter and receiver functions in a single unit to send and receive communications signals. In optical networks, a transceiver converts electrical signals into optical signals for transmission over fiber and receives optical signals back into electrical signals. Optical transceivers can use different transmission technologies, including intensity modulation with direct detection (IM-DD) and coherent detection.
transmission loss
The reduction in signal power that occurs as a signal travels through a transmission medium or communications system. In optical networks, transmission loss can result from fiber attenuation, connectors, splices, and other optical components and is typically expressed in decibels (dB).
U
Ultra Ethernet
An open, Ethernet-based communications architecture developed by the Ultra Ethernet Consortium (UEC) to meet the performance and scalability requirements of large AI and high-performance computing environments. It enhances Ethernet with capabilities for improved congestion control, load balancing, transport efficiency, and reliable, high-performance scale-out connectivity.
V
W
wavelength
The distance between successive corresponding points of a wave, such as two adjacent crests. In optical communications, wavelength identifies the optical carrier used to transmit information and is typically measured in nanometers (nm). In wavelength division multiplexing (WDM) systems, multiple wavelengths are used to carry independent optical channels over the same fiber.
wavelength division multiplexing (WDM)
An optical transmission technique that combines multiple independent optical signals, each carried on a different wavelength, for simultaneous transmission over a single optical fiber. WDM significantly increases fiber capacity by allowing many wavelength channels to share the same fiber.
wavelength selective switch (WSS)
A programmable optical device that independently directs or attenuates selected wavelength channels between input and output ports. WSS is a key building block of ROADMs, enabling individual wavelength channels to be routed and power-controlled without optical-electrical-optical (OEO) conversion.
X
eXtra-dense Pluggable Optics (XPO)
A high-density, liquid-cooled pluggable optical transceiver form factor designed to support higher optical interface density and power than conventional air-cooled pluggable form factors. By providing direct liquid cooling to the optical modules, higher-power transceivers can be packed more densely, supporting the increasing bandwidth requirements of AI and high-capacity data center networks.
Y
YANG
A data modeling language standardized by the Internet Engineering Task Force (IETF) for defining configuration data, operational state, remote procedure calls, and notifications associated with network devices and services. YANG defines the structure of the data, while protocols such as NETCONF and RESTCONF provide ways to access and manipulate it. APIs support broader software-to-software interaction, and software-defined networking (SDN) uses programmability and interfaces such as these to automate and control the network.
Z
zero-touch provisioning (ZTP)
The automated configuration and activation of a device or service with little or no manual intervention. When initially deployed, a device can automatically retrieve its software, configuration, and operational parameters from centralized systems, reducing deployment time, operational effort, and configuration errors.
ZR+
A general industry term for coherent pluggable optical interfaces that provide capabilities beyond those defined by basic ZR specifications. Unlike ZR, ZR+ does not refer to a single standardized specification. ZR+ implementations may support additional data rates, stronger forward error correction (FEC), enhanced optical performance, and longer-reach, multi-span transmission applications.