Computer Science Glossary

Finite and explicit sequence of steps that solves a problem or performs a computation; not necessarily code.

Qualifies a signal or system whose values vary continuously; its conversion to digital notably requires sampling.

Capacity of a channel to carry data per unit of time; the actual throughput obtained may be lower.

Base-two numeral system using 0 and 1; it is central to the representation of computer data.

Smallest unit of binary information, taking two conventional values, 0 or 1.

Measure of how the resources required by an algorithm grow, notably its runtime or memory.

Reduction of the volume of data to be stored or transmitted, with or without information loss depending on the method.

Compression that drops some information to reduce size further, without perfectly identical reconstruction.

Compression that allows the original data to be reconstructed exactly after decompression.

Number of bits transmitted per unit of time, usually expressed in bit/s and not in bytes per second.

Base-ten numeral system used in everyday life; it is distinct from the base-two and base-sixteen systems common in computing.

Operation that interprets or reverts an encoded representation according to the rules of the format used.

Recordable or transmittable representation of a fact, measurement or content, intended to be interpreted or processed.

Transformation of information according to representation rules, without necessarily implying confidentiality.

Measure of uncertainty or information quantity within a model; its definition depends on the statistical or cryptographic context.

Number of cycles per second of a periodic phenomenon; a processor's GHz alone do not determine its performance.

Decimal unit equal to one billion bytes; the gibibyte (GiB) equals 2^30 bytes.

Base-sixteen numeral system using 0-9 and A-F, convenient for compactly displaying binary values.

Information indicating a date and time associated with an event, often recorded in a standardised format.

Meaning obtained by interpreting data in a context; data alone does not always suffice to inform.

Discipline devoted to the automatic processing of information by machines, and to the design of the methods and systems that enable it.

Ability of distinct systems to exchange data and exploit it through compatible interfaces and conventions.

Binary unit equal to exactly 1,024 bytes, often confused with the commercial kilobyte.

Decimal unit of 1,000 bytes; to be distinguished from the kibibyte, which contains 1,024.

Delay between a request and the response or expected effect, distinct from the volume transferable per second.

Decimal unit equal to one million bytes; the mebibyte (MiB) equals 1,048,576 bytes.

Data describing other data, for example its date, author, size or format.

Group of four bits, often represented by a single hexadecimal digit.

Notation describing an asymptotic growth bound, often used to compare the costs of algorithms.

Qualifies information represented by discrete values, generally binary-encoded, as opposed to a continuous analog signal.

Group of eight bits; in modern systems 'byte' usually denotes an octet, but its historical meaning has varied.

Decimal unit equal to one trillion bytes; the tebibyte (TiB) equals 2^40 bytes.

Component that converts mains electricity into voltages suitable for computer components.

Part of the processor that performs numerical and logical operations such as additions or comparisons.

Family of processor architectures based on the ARM instruction set, widespread on mobile devices and also present in computers and servers.

Organisation of hardware and software components, and of the rules that determine how they work together.

Family of 64-bit instruction sets common on personal computers and servers.

Test measuring certain performance aspects of a device or software in a defined scenario, without summarising all uses.

Historical PC initialisation firmware; recent machines generally use UEFI firmware instead.

Communication pathway carrying data, addresses or signals between hardware components.

Processor cache levels with different capacities and speeds; their organisation varies by chip.

Main circuit board connecting processor, memory, storage and expansion interfaces of a computer.

Set of circuits or controller functions managing certain communications between components; part of it may be integrated into the processor.

Compute unit of a CPU capable of executing one stream of instructions; multiple cores enable more parallelism.

Component that removes heat from a part, possibly assisted by a fan or a liquid loop.

Non-volatile memory that is electrically erasable and reprogrammable, used notably for hardware parameters.

Software embedded in a device to initialise or control its functions, distinct from an ordinary application.

Resource that limits the overall performance of a system even when other resources are available.

Processor designed for many parallel computations, notably graphics rendering and certain scientific or AI operations.

Set of operations and conventions accessible to machine software, for example x86-64 or ARM.

Fast memory close to the processor that holds copies of data and instructions to limit slower accesses.

Memory intended to retain data without power, often used for system code; some variants are rewritable.

Generally volatile working memory in which programs store their data while running.

Accelerator specialised in certain AI model operations, in particular inference.

Programmable device that runs instructions to process, store and exchange data.

Setting a component beyond its nominal frequencies, with possible effects on heat, power draw and stability.

High-speed interconnect used notably by graphics cards, SSDs and expansion cards.

Component that executes the general instructions of programs and coordinates much of the computation.

Small storage area inside the processor used to manipulate values and addresses very quickly.

UEFI feature that verifies the signatures of boot components according to a configured trust policy.

Chip integrating multiple functions, for example CPU, GPU, controllers and interfaces, in a single package.

Automatic reduction of a component's performance or frequency to keep its temperature in check.

Execution context supported by a core, able to share its resources with other simultaneous contexts.

Hardware or integrated module providing cryptographic functions and key protection, used among other things for measured boot.

Modern firmware interface responsible for hardware initialisation and launching the operating system.

In common usage, the case that houses the main components of a desktop computer; not to be confused with the CPU.

Memory used by a graphics processor for textures, images, buffers and other data needed by its computations.

Short-range radio technology used notably for accessories, audio and exchanges between devices.

Input device for text and commands; its layout may be AZERTY, QWERTY or another.

Number of physical pixels that can be displayed in width and height, distinct from size and density.

Digital display interface designed to carry image and optionally sound and auxiliary data.

Output device that shows images and text, whose characteristics include definition, size, brightness and refresh rate.

Digital audio/video interface commonly used between computers, monitors and televisions.

Technique that allows representing a wider range of brightness and colours when display, format and content support it.

Output device producing on paper or another medium a physical representation of a digital document.

Sensor that converts sound vibrations into a signal usable by an audio interface.

Very short-range radio communication used for example for pairing and some contactless payments.

Surface sensitive to finger movements, used to move the pointer and perform gestures.

Hardware connected to a computer to provide input, output, storage or communication.

Hardware that provides information to the system, such as a keyboard, mouse or microphone.

Hardware that returns information, such as a screen, printer or speakers.

Software that allows the operating system to communicate with a particular component or peripheral.

Mechanism that facilitates automatic detection and configuration of a compatible peripheral.

Pixel density per unit of length; everyday usage sometimes confuses density with definition.

Device that digitises a physical document as an image, possibly converted to text by optical recognition.

Pointing device that moves a cursor and sends actions through buttons or a wheel.

Accessory that gathers several connections to quickly link a computer to displays and peripherals.

Number of display updates per second, expressed in hertz; it differs from the number of frames rendered per second.

Time required for a change of pixel state, distinct from the overall display latency.

Interconnect technology carrying data and, depending on versions and equipment, video and power.

Family of connection and communication standards for peripherals; plug type and protocol performance are not synonymous.

Reversible connector form that can carry different USB versions or other protocols depending on the device.

Camera designed to transmit or record digital images, notably for video conferencing.

Apple file system designed for modern media, featuring functions such as clones and snapshots.

Long-term preservation of data for its documentary or regulatory value, distinct from the rapid recovery provided by a backup.

Linux file system integrating notably snapshots, checksums and advanced volume management.

Family of removable storage media using flash memory, with several sizes and speed classes.

Portable flash storage medium using a USB interface, with varying capacities and performance.

Technique that stores identical blocks or objects only once to reduce the space used.

Magnetic platter-based storage medium, generally slower than an SSD but economical for high capacities.

Medium read or written by laser, notably CD, DVD or Blu-ray, now mainly used for media and archives.

File system designed for removable media and large files, with broad compatibility between devices.

Journaling file system common on Linux, suited to general-purpose volumes.

Old, widely compatible file system, but limited for example to files smaller than 4 GiB.

Preparation of a medium or volume with a usable organisation, often a file system; it does not guarantee secure erasure.

Logical representation of the state of a volume or system at a given time; it is not automatically an independent backup.

Physical card and connector form factor that can host different interfaces, including SATA or PCIe depending on the device.

Old boot and partitioning scheme whose limitations explain the frequent move to GPT.

Non-volatile, electrically erasable and rewritable memory technology used notably in SSDs, USB drives and memory cards.

Network-attached storage device that allows several machines to share data.

File system primarily used by Windows, supporting notably permissions and journaling.

Protocol optimised for accessing non-volatile storage, often over PCI Express; it is not a physical form factor.

Logical division of a storage medium into named zones that the system can manage separately.

Organisation of several disks aiming, depending on the level, at performance, availability or both; it never replaces a backup.

Backup principle aiming at three copies, two types of media and one offsite copy; other protections may be required.

Operation that recovers data or a system from a backup, snapshot or emergency medium.

Specialised network presenting block storage resources to servers, different from the file sharing typical of a NAS.

Link interface notably used by hard drives and SSDs; it is distinct from the PCIe interfaces used by many NVMe SSDs.

Copy of data that allows restoring it after loss, corruption or incident; it must be tested and isolated according to risk.

Copy of the entire selected perimeter at a given time, possibly used as a reference for other backups.

Copy of changes made since the last full backup, often simpler to restore than a long incremental chain.

Copy of changes made since the previous backup in the considered chain.

Storage medium mainly using flash memory, with no moving parts to read or write data.

Retention of data on a medium, generally non-volatile, so it can be retrieved after a program finishes or shuts down.

Rules and structures used to record, retrieve and organise files and directories on a volume.

Modern partition description scheme commonly associated with UEFI systems, distinct from the historical MBR format.

Logical storage space presented to the file system, possibly based on one or more partitions or physical units.

Request a program addresses to the kernel for a privileged operation, such as opening a file or creating a process.

Controlled procedure that ends services, synchronises data and prepares the machine for power-off.

Program responsible for selecting or launching the operating system after the first firmware steps.

Organisation of several tasks whose execution periods overlap, without requiring them to physically run at the same instant.

Sequence from power-on to hardware initialisation and loading of the operating system.

Coherent set combining a kernel, system tools, package manager and configuration choices, for example Fedora or Debian.

Extended privileges allowing modification of protected settings and resources; 'root' denotes the superuser on Unix/Linux.

Privileged execution context in which the kernel and certain system components run.

Execution area of applications normally holding fewer privileges than the kernel.

Stream of instructions of a process able to share its memory with other threads of the same process.

Persistent consumption of memory that has become unused but is not freed or is wrongly retained by a program.

Sleep mode that saves the system state to non-volatile storage before cutting main power.

Signal inviting the processor to temporarily pause its work to process a hardware event.

Recording of system and application events useful for diagnostics and audit, subject to protecting the data they contain.

Family of systems built around the Linux kernel; a distribution adds notably tools, libraries and package management.

Apple's operating system for Mac computers, based notably on Darwin and Apple-specific components.

Mechanism providing processes with an abstract address space, mapped onto physical memory and possibly storage.

Use of several distinct processes to perform tasks, notably by leveraging multiple cores.

Ability of a system to advance several tasks by sharing processor time or parallel execution.

Use of several threads in a program, able to improve responsiveness or parallelism depending on the work.

Central component of the operating system responsible notably for processes, memory, drivers and hardware access.

Component deciding which processes or threads can use computing units and for how long.

Memory management in fixed-size blocks called pages, used for addressing and loading.

Effective execution of operations at the same time, notably on several cores or computing units.

Environment variable indicating the directories where to search for executable commands when their full path is not provided.

Set of standards favouring the portability of programs and interfaces between Unix-like systems.

Instance of a running program, having its own context and resources managed by the system.

Background program providing a function to other processes, e.g. logging or networking.

Storage area the system can use to move certain memory pages out of RAM, with a performance cost.

Fundamental software managing hardware resources and providing services to programs and users.

Windows function requesting elevation or validation for actions requiring elevated privileges.

Family and tradition of operating systems influencing many current environments; UNIX is also a certification mark.

Named value passed to processes to configure them, for example the search path for executables.

State that reduces power consumption while allowing activity to resume quickly, depending on hardware and chosen mode.

Family of Microsoft operating systems for computers and other categories of machines.

Software intended to perform one or more tasks for a user or another system.

Application combining web technologies with a container or native functions of the platform.

Software developed to take direct advantage of the interfaces and conventions of a given platform.

Application used through a browser and web technologies, with variable local capabilities depending on its design.

Set of functions or resources reusable by several programs.

Ability of a new version to support content or uses from earlier versions, within a defined scope.

Free-software approach that imposes, under certain redistribution conditions, the maintenance of similar freedoms and obligations for derived versions.

Targeted modification intended to fix a defect or vulnerability, sometimes delivered as an update.

Modification distributed quickly to resolve an incident considered a priority.

Library, module or program required for another component to function.

Organised location publishing packages and metadata needed for their discovery and installation.

Date or state from which the vendor ceases certain announced forms of support or maintenance.

Model offering a free base offering and additional paying features, quotas or services.

Software distributed free of charge, without that implying access to source code or free-software freedoms.

Tool that installs, updates and removes software while managing dependencies and repositories.

Program or package responsible for placing an application's files and preparing its configuration on a device.

Contract setting out notably the conditions of use, modification and redistribution of the software.

Set of programs, data and resources that performs a function on a computer system.

Software granting the freedoms to use, study, modify and redistribute its code, according to the conditions of its licence.

Application designed to run without traditional installation or with few host-system dependencies.

Software whose use, modification and redistribution depend on rights granted by its owner; its code is not necessarily accessible.

Program that provides or administers base technical services rather than a direct business task.

Release bringing fixes, improvements or functions to a program or a system.

Qualification of software distributed under a recognised licence allowing access, modification and redistribution of code; not guaranteed free of charge.

Ability of software to be adapted or executed on different systems with limited effort.

Web application that may be installable and provide certain advanced features, possibly offline depending on the browser and implementation.

Defect introduced by a change that degrades a previously functional behaviour.

Candidate version intended for final validation before stable release, barring problems discovered.

Set of components required to execute software, for example a virtual machine, libraries or a language engine.

Software delivered as a network-accessible service, generally operated by its vendor, with varying contractual terms.

Software offered for trial or with restrictions before purchase or full licensing.

Identifier distinguishing a software release, according to a convention specific to the project.

Early development release whose interface, functions and stability may still change significantly.

Release proposed for wider testing before stabilisation, without guaranteeing the absence of defects.

MAJOR.MINOR.PATCH versioning convention linking announced changes to the compatibility of the public interface.

List of rights associated with identities or groups to control access to a resource.

Hierarchical organisation of folders and files linked by parent-child relationships.

File grouping several items and their metadata; an archive is not necessarily compressed.

Old 7-bit encoding representing 128 characters and control codes, mainly suited to English.

Optional signature that may indicate byte order for some Unicode encodings; its use in UTF-8 is generally not required.

Address of a file or folder described from a fixed starting point such as the system root.

Address interpreted relative to a reference directory rather than the root.

Logical place where the system moves certain deleted files to ease recovery, without guarantee of secure erasure.

Tabular text data format with separators; variants of delimiter, quoting and encoding sometimes require settings.

Logical container grouping files and other directories in a tree structure.

Convention linking characters to sequences of bytes or code units, essential to correctly reading text.

Suffix of the name, often after a dot, suggesting a format without guaranteeing the file's actual content.

Set of data identified and managed as a unit by a file system.

File whose bytes are not necessarily interpretable as text according to a given encoding.

File whose display is hidden by default according to system conventions, without intrinsic security protection.

File that can be launched as a program in a compatible environment and with the required permissions.

File created for intermediate processing and normally deleted when no longer useful.

File containing a representation of characters interpretable using an encoding such as UTF-8.

Sequence marking line separation: LF is common on Unix, CRLF on Windows and in some protocols.

Rules defining the structure and interpretation of a file's data, independently of its visible extension.

Text format structuring objects, arrays and values, used for data exchange and configuration.

Additional name referencing the same file object on a compatible system, not a mere path shortcut.

File-system entry that points to another path and may become invalid if its target is moved.

Lightweight markup syntax for writing structured documents in plain text.

Document format aiming for stable presentation across environments, with capabilities for text, images, links and forms.

Rule determining who can read, modify or execute an item; models vary between systems.

Starting point of a file tree, represented by '/' on Unix-like systems.

Standardised identifier indicating the nature of a content, for example text/plain or image/png.

Standard assigning code points to characters of many languages, distinct from its encodings such as UTF-8.

Unicode encoding using one or two 16-bit units to represent a code point.

Variable-length Unicode encoding using one to four bytes per code point, very common on the Web.

Extensible markup language representing documents and structured data through elements and attributes.

Readable serialisation format often used for configuration, sensitive notably to indentation and certain type conventions.

Design of products usable by people with varied capabilities, notably with assistive technologies.

Desktop area giving access to applications, windows and sometimes system notifications and controls.

Main graphical space of the system presenting windows, shortcuts and sometimes files.

Bar grouping applications or documents to facilitate their launch and navigation, depending on the system.

Overall quality experienced when using a product, including understanding, efficiency, accessibility and satisfaction.

Application that browses, organises, copies, moves and deletes a system's files.

Display area of an application that can be moved, resized or stacked according to the interface manager.

Gesture that moves an element with a pointer to trigger an action in another area.

Interface in which operations are performed by typing textual commands.

Interface based on visual elements such as windows, buttons, icons and menus.

Set of elements through which a person interacts with software or a device.

Assistive software that renders interface-accessible information through speech or braille.

Approach favouring data and interactions on the user's device, with possible synchronisation without permanent server dependency.

Menu presenting actions relevant to the selected item or pointed area.

Operation that does not require a network connection at a given time, depending on actually available functions.

Interface colour scheme favouring dark backgrounds; it does not by itself guarantee readability or accessibility.

Ability to browse and use an interface without a mouse or touch command.

Message emitted by software to signal an event without requiring the main screen to be opened immediately.

Technique that transforms the image of printed or handwritten characters into usable text, with variable accuracy.

Element grouping several views or documents within the same window.

Document organised in slides intended to support a visual and spoken communication.

Temporary storage used to copy or cut content before pasting it elsewhere.

Combination of keys that triggers a command without going through menus.

Bringing copies or states of data into agreement between devices and services, sometimes with conflict handling.

Software manipulating data in cells, formulas and tables for calculation, analysis and representation.

Software dedicated to writing and laying out textual documents, with various editorial functions.

Logical identifier assigned to an interface or device on a network using the IP protocol.

Address reserved for certain networks not directly routed on the public Internet; its range depends on IPv4 or IPv6.

Address that can be routed on the Internet subject to network policies and proper connectivity operation.

Link-layer identifier associated with a network interface, modifiable or randomised in certain contexts.

Device provided or preconfigured for Internet access, possibly combining modem, router, Wi-Fi and telephony.

Notation and method for grouping IP addresses using a prefix length, for example /24.

Device connecting devices on a local network and mainly forwarding frames according to their link-layer addresses.

Legacy Ethernet device that repeats signals to several ports, without the selection comparable to that of a switch.

Protocol that automatically assigns network parameters, such as IP address, gateway and DNS servers.

Distributed system associating domain names with various pieces of information, notably IP addresses.

Family of local area network technologies and protocols, mainly used on wired links.

Variation in transmission delay between packets, particularly disruptive for live voice and video.

Version of IP using 32-bit addresses, frequently associated with address translation due to insufficient public address space.

Version of IP using 128-bit addresses, designed notably to greatly expand the address space.

Network covering a limited space such as a home, an office or a corporate site.

Value separating the network part from the host part of an IPv4 address, generally also expressed as a CIDR prefix.

Device that modulates and demodulates a signal adapted to an access medium, depending on the connection technology.

Maximum size of a packet or transmission unit on a given link without fragmentation at the layer concerned.

Network address translation modifying the addresses of certain packets, common for sharing an IPv4 connectivity.

Element connected to the network, such as a server, workstation, router or communications device.

Unit of data transmitted at a network layer; its content may be encapsulated in frames or segments.

System filtering communications according to rules; it can be software, hardware or provided as a service.

Router to which a device sends packets destined for networks it does not know directly.

Proportion of packets that fail to reach their destination, potentially disrupting real-time communications.

Tool that checks reachability and estimates a round-trip time via ICMP messages, whose filtering may bias interpretation.

Device providing a wireless connection to a network, sometimes integrated into a home router.

Logical number identifying a TCP or UDP communication endpoint on a device, distinct from a physical socket.

Intermediary that receives then forwards requests between a client and a service, for control, caching or confidentiality depending on use.

Traffic priority and management mechanisms that influence delay, throughput or losses according to needs.

Set of interconnected devices that exchange data following agreed protocols.

Device or software that forwards packets between IP networks according to a routing table and rules.

Software endpoint of communication between processes, local or networked, described by an address family and protocol.

Name identifying a Wi-Fi network; this name alone is not a security measure.

Transport protocol establishing a reliable, ordered stream between two endpoints, with retransmission and congestion control.

Physical or logical organisation of connections between nodes, for example star or mesh.

Tool showing part of the path taken by packets through routers, with visibility depending on network responses.

Link-layer transmission unit, including control and addressing information specific to the protocol.

Datagram-oriented transport protocol that guarantees neither delivery nor order, useful for applications handling these properties themselves.

Logical separation of networks at the link layer on a shared infrastructure, without alone guaranteeing their security.

Logical tunnel between networks or devices possibly protecting communication with encryption; its guarantees depend on configuration.

Network linking remote sites, often through operator infrastructures.

Family of wireless local area network technologies based on IEEE 802.11 standards.

Local area network using a radio link, frequently implemented with Wi-Fi.

Interface accessible via network protocols allowing software to exchange data or perform operations.

Entity that issues and signs digital certificates according to a certification policy.

Local temporary storage of web resources intended to speed up visits and limit transfers.

Email copy fields: CC recipients are visible, BCC recipients are normally not visible to the other recipients.

Distributed infrastructure that brings some resources closer to users to reduce load and access time.

Electronic document linking an identity or name to a public key, according to a chain of trust and validation rules.

Small data stored by the browser for a site, notably to manage a session or preferences.

Electronic message exchange service between addresses, generally passing through mail servers.

Company or organisation giving access to the Internet to subscribers via wired, radio or satellite infrastructure.

Query language and runtime environment that lets the client request exactly the fields it needs in an API schema.

Provision of technical resources intended to make a site or application accessible on a network.

Request and response protocol used to transfer web resources between clients and servers.

HTTP used over a connection protected by TLS, providing notably encryption and authentication depending on certificate validation.

Element allowing navigation from one digital resource to another using an address.

Protocol allowing messages to be consulted and managed on a server while synchronising their states.

Global infrastructure interconnecting networks based mainly on the IP protocol suite.

Service indexing content and offering results in response to user queries.

Software that retrieves, interprets and displays web resources, while possibly executing client-side code.

Hierarchical name associated with a DNS zone, for example example.org, distinct from the site hosting.

Document or view accessible on the Web, generally in HTML, possibly including styles, scripts and external content.

File or content attached to an email, subject to size constraints and possibly presenting security risks.

Protocol allowing retrieval of emails from a remote mailbox, generally with fewer synchronisation features than IMAP.

Architectural style defining constraints for distributed systems; an HTTP API is not automatically REST.

Software and, by extension, machine that responds to HTTP requests by providing pages, files or application responses.

Logical state linking several exchanges of a user with a service, often associated with a cookie or token.

Organised set of web pages or resources associated with a common identity, domain or function.

Protocol mainly used for sending and transferring emails between clients and servers or between servers.

Unsolicited electronic message sent in bulk, often advertising or fraudulent.

Transmission allowing playback of content to start before its full download, with various buffering strategies.

Transfer of data from a remote system to a local device; transfer in the other direction is often called upload.

Part at the top of the domain name hierarchy, such as .fr or .org.

Cryptographic protocol protecting application exchanges, notably those of the Web in HTTPS.

Sending of data from a device to a remote server or service.

Resource identifier that may take various forms; a URL is a category of URI.

Address describing the location or mode of access to a resource, generally composed of a scheme, a host and a path.

Notification automatically sent by a service to a URL when a configured event occurs.

Protocol enabling persistent bidirectional communication after the connection is established.

Service for consulting linked resources identified by URLs, based notably on HTTP; it is only one service of the Internet.

Hierarchical structure composed of nodes linked by parent-child relationships.

Logical value normally taking one of two states: true or false.

Construct repeating a block of instructions according to a counter, traversal or stop condition.

Sequence of characters intended to represent text, with encoding and operations depending on the language.

Model used to define or create objects in certain object-oriented languages.

Original text or representation of a program from which it is executed, compiled or assembled.

Construct selecting a processing branch based on the result of a logical test.

Value or binding the program does not allow to modify within a defined scope; the notion varies by language.

Structure associating keys with values to retrieve a datum from its key.

Organisation that groups state and behaviour and controls the means of access to internal details.

Collection whose each element is unique according to an equality rule, generally without guaranteed positional indexing.

Numeric type representing numbers without fractional part, with a range depending on its implementation.

Mechanism signalling an abnormal or particular event, able to alter the normal execution flow.

Structure where the first element added is generally the first removed, called FIFO.

Reusable code block that performs a processing, optionally with parameters and a result.

Mechanism that automatically frees part of memory that has become unreachable, without eliminating all logical leaks.

Set of strategies that detect, signal and handle failures without confusing all anomalies.

Set of vertices and edges or arcs used to represent complex relationships.

Mechanism allowing a type to reuse or specialise the characteristics of another, depending on the language.

Communication contract exposing functions, data or operations usable by other components.

Repetition of a processing or successive traversal of the elements of a collection.

Formal language used to describe computations, processes and the behaviour of a program.

Structure whose elements are linked by references, enabling certain insertions without large data movement.

Approximate representation of real numbers on a finite number of bits, subject to rounding errors.

Software entity grouping state and behaviour according to an object-oriented programming model.

Name of a datum declared in a function; the argument is the value passed at call time.

Data structure where the last element added is the first removed, called LIFO.

Value or object containing the address of a memory area, with usage rules depending on the language.

Ability of an interface or operation to apply to values of different types under common rules.

Region of the program in which an identifier is accessible and where its meaning is determined.

Activity consisting of designing and writing instructions that the computer can execute or interpret.

Technique in which a function or definition calls itself, directly or indirectly.

Way to designate a datum or object without necessarily copying it, depending on the language's guarantees.

Meaning attributed to the constructs of a language and to the effects of their execution.

Rules that determine the constructs allowed in a language, distinct from their meaning.

Ordered collection of elements generally accessible by index; its memory layout depends on the language.

Classification determining which values and operations are allowed, for example numbers, text or booleans.

Result that a function hands back to its caller, distinct from the side effects it may produce.

Representation of the absence of value or reference in some languages, distinct from zero or the empty string.

Name or logical location associated with a value that can be consulted and sometimes modified during a computation.

Examination of code without executing it to detect potential errors, rule violations or type incompatibilities.

File produced by the build, such as a binary, a package, a container image or a report.

Movable reference to a line of development allowing work on distinct changes.

Production of software artefacts from sources and dependencies, often through compilation, assembly and packaging.

Intermediate instruction representation intended to be executed or compiled by a virtual machine or a runtime.

Automation of version preparation and, depending on the practice chosen, of its release to production.

Practice that regularly runs build, analysis and tests on every integrated or proposed change.

Recording of a state and its metadata in the Git history; it does not imply pushing to a server.

Tool that performs compilation of a program from a source language to a target representation.

Transformation of code to a target representation, for example a native executable or bytecode, depending on the tool.

Situation where modifications cannot be combined automatically and require a decision.

Indicator of the share of code or scenarios exercised by tests; a high rate does not guarantee their quality.

Tool helping to observe execution, set breakpoints and inspect values to find defects.

Future cost generated by design or maintenance compromises, voluntary or not.

Information explaining architecture, interfaces, deployment or maintenance of a system, intended to ease its use and upkeep.

Tool for entering and modifying source text, often enriched with highlighting, completion and navigation.

Transformation of code presentation according to a convention, usually without altering its behaviour.

Software structure imposing or facilitating an application organisation and providing reusable components.

Organised tracking of a project's modifications to preserve its history and ease collaboration.

Distributed version control system used to record change history and coordinate work.

Software combining code editing, build tools, debugging and other functions aimed at developers.

Software executing instructions of a language or intermediate representation without systematically requiring a native binary.

Static analysis flagging style, consistency or certain probable defects in code.

Software environment that executes a program representation, for example bytecode instructions.

Operation that combines histories or branch changes, possibly requiring conflict resolution.

Automated sequence of validation, build and delivery steps for software.

Instruction or location where a debugger suspends execution to allow inspection.

Measurement of CPU, memory or other resource usage to identify execution costs.

Proposal of changes to be reviewed and integrated into a target branch via a collaborative platform.

Git operation that reapplies commits onto a new base and changes their identifiers.

Internal reorganisation of code while preserving its intended observable behaviour, to improve its structure.

Program often used to automate operations and executed by an interpreter or dedicated environment.

Set of tools, libraries and documentation provided to develop on a platform or with a service.

Program interpreting commands or scripts, for example Bash, Zsh or PowerShell.

Textual interface giving access to a command environment and the output of launched programs.

Test reproducing a complete journey through several layers of the application, often close to real usage.

Test evaluating a system under a defined volume of users, requests or transactions.

Test intended to detect the reappearance of a defect or a degradation after a software modification.

Test verifying the communication and joint behaviour of several components or services.

Automated test targeting a limited component and its expected behaviour under controlled conditions.

Structured recording of the path taken by operations to understand their progress and performance.

Type checking mainly during operation execution, without implying the absence of types.

Checking of part of the type constraints before execution, the extent of which depends on the language and tooling.

Organisation of components reacting to published events rather than to systematic direct calls.

Organisation of an application's components, their responsibilities and their exchanges to meet technical and business constraints.

Separation of presentation, business logic and data layers; logic and deployment do not always coincide.

Organisation allowing an operation to be launched and its response handled later, without systematically blocking the calling flow.

Strategy that progressively spaces new retries, often with a random component to avoid simultaneous spikes.

Infrastructure allowing services to exchange messages using publication or distribution mechanisms.

Temporary storage of computed or loaded results to reduce repeated computations and accesses.

Function provided to a component to be invoked when an event occurs or a processing finishes.

Model in which clients request resources or services from servers, possibly distributed across several machines.

Degree to which the elements of a component contribute to a clear and common responsibility.

Degree of dependency between components; loose coupling often facilitates their independent evolution.

Model separating read and modification operations, sometimes down to data models or infrastructures.

Reconstruction of structured data from a representation; untrusted formats require special precautions.

System that stores messages destined for consumers in order to decouple production from processing.

Remote procedure call framework notably relying on HTTP/2 and Protocol Buffers in its common uses.

Property of an operation whose repetition produces the same intended effect as a single execution, with comparable initial state.

Mechanism that removes or refreshes cached data when it becomes stale.

Architecture composed of autonomous services exchanging through interfaces, with extra coordination and operation costs.

Intermediate component providing functions such as authentication, logging or request transformation.

Division of a system into elements with identified responsibilities, connected through controlled interfaces.

Application whose many features are developed and delivered as a mainly unified whole.

Splitting a result set into pages accessible by index, offset or cursor.

Object representing the future result or failure of an asynchronous operation in some languages.

Format and schema system for serialising structured messages in a compact and interoperable way.

Model in which producers publish on topics and subscribers receive the corresponding events.

Rule capping the frequency or volume of requests accepted by a service.

Controlled re-execution of an operation after failure, with safeguards against duplicates and contention.

Mechanism invoking a remote procedure as a structured operation, with network costs and failures to manage.

Conversion of a data structure into a transportable or storable representation, such as JSON or binary.

Time limit beyond which an operation is abandoned or flagged as failed.

Set of properties expected of transactions: atomicity, consistency, isolation and durability, depending on the system considered.

Exploration, preparation and interpretation of data sets to produce information or inform decisions.

Property of a transaction whose operations are either committed together or rolled back together.

Organised set of data whose access and updates follow defined rules.

Term grouping various databases not centred on the classic relational model, notably document, key-value and graph.

Database organising data as relations, often presented in tables and queried with SQL.

System designed to store vectors and search nearest neighbours according to a similarity measure.

Set of issues linked to data whose volume, velocity or variety exceed ordinary available tools.

Column or set of columns referencing a key of another table or of the same table to enforce a relationship.

Column or set of columns that uniquely identifies the rows of a table under a declared constraint.

Property defined in a table schema that receives a value for each row.

Storage gathering raw or lightly transformed data of various formats for later uses.

Infrastructure centralising and structuring data for analysis and reporting.

Property by which a committed transaction remains recorded despite certain incidents covered by the system.

Extract, transform and load process moving data between systems for exploitation.

Additional structure accelerating certain searches at the cost of space and modification costs.

Structure associating to each term the list of documents in which it appears, speeding up textual search.

Level of separation between concurrent transactions, influencing possible anomalies and synchronisation cost.

Operation combining rows from several tables according to a matching condition.

Set of values representing a data occurrence in a relational table.

Controlled modification of a database's structure, sometimes needing data and code adjustments.

System indexing resources within a given perimeter to find those matching a query.

Splitting a data set into partitions to ease management, storage or performance.

Search operating on the indexed textual content of documents rather than only on their metadata.

Search aiming at proximity of meaning between query and documents, often through vector representations.

Copy of data or changes to one or more instances for availability, reading or recovery.

Instruction requesting to consult or modify data according to defined syntax and criteria.

Return of a transaction's uncommitted modifications to a consistent earlier state.

Description of the structure and constraints of the data, often including tables, types, relations and indexes.

Database management system providing storage, queries, access control and consistency according to its capabilities.

Horizontal distribution of data across several nodes or databases, often by a distribution key.

Standardised language used to define, query and modify relational data, with dialects depending on the DBMS.

Relational structure containing rows organised according to columns defined by a schema.

Set of operations executed as a logical unit, with guarantees depending on the engine and isolation level.

Graphical representation of data allowing trends, distributions and relationships to be read without replacing critical analysis.

Multiplication of secret or value attempts to discover an identifier, key or password.

Verification of the claimed identity of a user, service or device; it generally precedes authorisation.

Decision determining which actions an authenticated identity or subject is allowed to perform.

Program presenting itself as legitimate while containing or facilitating a hidden malicious action.

Transformation of data using an algorithm and a key to render its content incomprehensible without a means of decryption.

Cryptography using a public and private key pair for certain encryption or signature operations.

Encryption using the same secret key, or closely related keys, to encrypt and decrypt.

Secret part of a cryptographic pair, intended to be protected against disclosure and unauthorised use.

Diffusable part of a cryptographic pair, used according to the algorithm to verify a signature or encrypt.

Property ensuring that only authorised people or entities can access information.

Set of mechanisms deciding who may access which resources and under which conditions.

Attack that makes the authenticated browser send an unintended action if the service lacks proper protections.

Set of practices, technologies and organisations aimed at protecting systems, services and data against digital incidents and attacks.

Distributed denial-of-service attack aiming to make a resource unavailable by overload or exhaustion.

Property by which a resource or service remains accessible when it is needed and authorised.

Technique or code taking advantage of a vulnerability to obtain an unintended effect; its nature varies with the targeted flaw.

Function transforming a datum into a generally fixed-size digest, which is not reversible encryption.

Organisation of storage, access, rotation and revocation of passwords, tokens and sensitive keys.

Process of detection, qualification, containment, resolution and feedback after a security event.

Tool that stores and generates credentials in a protected vault, with filling features depending on the solution.

Storage of a derivative resistant to password computation, generally with a salt and a slow algorithm.

Fraud by deceptive messages or pages aiming to obtain secrets, payments or actions from the victim.

Discipline and tooling that administers an organisation's accounts, authentication, roles and permissions.

Action of declaring or determining a subject's identity, for example through an account name.

Manipulation of people to bypass protections or obtain disclosure of information and access.

Vulnerability where untrusted data is interpreted as SQL commands, allowing requests to be hijacked.

Property ensuring that data and systems are not altered in an unauthorised or undetected way.

Element presented to a service to exercise granted rights, with scope and duration depending on its issuance mechanism.

Compact token format containing claims signed or encrypted according to its profile; its presence does not guarantee confidentiality.

Software that covertly collects information or monitors activities without valid consent.

Software designed to harm, defraud or act without authorisation, including notably ransomware and spyware.

Verification combining at least two factors of different categories, such as a secret and a possession, to strengthen security.

Principle granting each user or process only the rights necessary for its mission.

Memorised secret used to verify access; its security depends notably on uniqueness, length and server-side protection.

Authorisation delegation framework allowing a client to obtain limited access to a resource without receiving the user's password.

Identity layer built on OAuth 2.0 enabling authentication and exchange of user information.

Protection filtering web requests according to rules targeting certain malicious behaviours, without replacing secure code.

Cryptographic authentication credential based on public keys, often protected by device unlock.

Authorised assessment simulating certain attacks to identify and demonstrate weaknesses within a defined scope.

Malicious software that blocks data or systems, often by encryption, to demand a ransom.

Value generally random and distinct per secret, added before derivation to limit precomputed-table attacks.

Cryptographic mechanism associating a verifiable signature with a message to control origin and integrity according to the keys used.

Team or structure monitoring security events, investigating alerts and coordinating responses.

Mechanism allowing access to several applications through central authentication, according to defined policies.

Set of entry and exposure points an adversary could attempt to exploit.

Malicious program that spreads autonomously between vulnerable systems or environments.

Malicious program able to infect other files or programs to replicate, often with host intervention.

Exploitable weakness of a component, system or process that may compromise its security.

Vulnerability allowing execution of unintended scripts in the context of a site visited by a victim.

Security model that avoids granting implicit trust based on network location and continuously verifies access according to context.

Configuration and deployment automation tool often based on inventories and declarative playbooks.

Site hosting servers, network, storage, power and cooling needed for digital services.

On-demand provision of computing resources accessible over a network, with pooling and varying billing terms.

Organisation combining private infrastructures with public cloud services through coordination or exchange mechanisms.

Cloud infrastructure reserved for a single organisation, hosted internally or by a provider.

Infrastructure operated by a provider and offered to several customers, with isolation mechanisms between them.

Set of machines or instances cooperating to provide common capacities or a distributed service.

Environment isolating an application and its dependencies while normally sharing the host system kernel.

Set of tools and technologies for building, distributing and running software containers.

Tool describing and launching several containerised services through a common configuration.

File declaring the instructions to build a Docker or compatible container image.

Execution model triggering functions on demand on a provider-managed infrastructure, often based on events.

Design aiming to keep a service accessible despite certain failures through redundancy and failover.

Software or control layer that creates and manages virtual machines on a host infrastructure.

Offering providing virtualised infrastructure resources such as compute, network and storage, partly administered by the customer.

Set of layers and metadata serving as an immutable template from which containers are created.

Management of infrastructure resources through declarative or programmable files that are versionable and automatable.

Container workload orchestration platform providing notably deployment, service discovery and declarative management.

Isolated computing environment with virtual hardware and generally its own guest operating system.

Adding or removing service instances to adjust capacity without merely increasing the power of each one.

Increasing or decreasing resources allocated to a given machine or instance.

Use of services from several cloud providers, which does not automatically imply interoperability.

Physical or virtual machine running workloads managed by a Kubernetes cluster.

Automation of deployment, scaling and repair of container sets.

Offering providing a platform to develop and run applications without managing all the underlying infrastructure.

Smallest common deployable unit of Kubernetes, grouping one or more containers that share certain contexts.

Geographic or administrative area in which a provider groups compute, network and storage resources.

Service storing and distributing container images, with possible access management and signatures.

Distribution of requests across several instances based on availability and routing rules.

Model in which the user manages servers less directly, even though physical infrastructure remains necessary.

System or machine providing the resources on which virtualised or containerised environments run.

Operating system executed in a virtual machine on the resources provided by the host.

Infrastructure-as-code tool describing resources and calculating changes to apply via providers.

Technique that exposes abstract computing resources from underlying hardware or systems.

Set of infrastructures designed to isolate certain types of failure within a cloud region.

Configuration, supervision and troubleshooting of equipment and services ensuring connectivity of an infrastructure.

Management of users, software, updates, resources and policies of a computing environment.

Signal emitted when a monitored condition requires analysis or intervention, with a risk of false positives.

Switch from a failing component to a backup resource to restore or preserve service.

Forecasting the resources needed from load, performance targets and observed trends.

Strategy using two comparable environments to switch traffic from the old to the new.

Progressive release of a version to a limited share of traffic to observe its behaviour before full rollout.

Set of practices bringing development and operations closer to automate delivery and improve service reliability.

Share of time during which a service meets the defined operating criterion over a given period.

Tracking the settings and states of equipment or software to ensure consistency, traceability and reproducibility.

Process that evaluates, approves and tracks changes likely to affect production.

Structured trace of important operations intended to reconstruct actions, with protection of its integrity and access.

Actions taken to resolve an observed failure and restore expected operation.

Planned interventions intended to reduce the probability of future failure or degradation.

Numerical measure of an aspect of a system, for example response time or memory occupancy.

Deployment of a version or configuration into the environment used by real users.

Collection and observation of indicators and events to detect problems and track a system's state.

Ability to understand a system's internal state from its traces, metrics and logs, depending on context.

Organisation of resources to maintain essential activities during a major disruption.

Component whose failure can cause the whole to stop because no replacement exists.

Procedures and resources planned to restore services after a disaster or major outage.

Duplication of critical resources or paths to reduce the impact of an isolated failure.

Controlled return to a previous version or configuration after an incident, subject to data compatibility.

Maximum tolerated data loss expressed as duration between the last recoverable state and the incident.

Maximum target time to restore a service after an incident, according to continuity requirements.

Contractual commitment describing a service level and possible consequences of non-compliance.

Measurable indicator used to quantify an aspect of the service, for example the percentage of successful requests.

Internal or shared service-level objective measured using precise indicators.

Discipline applying software engineering methods to operations to measure and improve reliability.

Duration of continuous operation of a device or service, an indicator different from all dimensions of its quality.

System using a model and control logic to plan or chain certain actions with tools, under defined conditions and limits.

AI approach that uses data to adjust a model making predictions or decisions on new cases.

Methods that look for structures in data without using explicit targets for the considered task.

Framework in which an agent learns behaviours from rewards and interactions with an environment.

Family of learning methods based on neural networks with several processing layers.

Training on examples accompanied by known targets, such as a label or value to predict.

Systematic distortion of results that may come notably from data, objectives, technical choices or uses of a system.

Technique that trains a model to reproduce part of another model's behaviour, often to obtain a lighter version.

Examples used to adjust a model's parameters during the learning phase.

Sample kept to evaluate the model's behaviour without having used it for fitting or selection.

Distinct data used notably to tune design choices and select a model during development.

Transformation of an element into a numerical vector intended to represent certain of its characteristics or relations.

Protocol and test set aimed at measuring certain capabilities or limits of a model, with risks of overfitting to the tests.

Maximum quantity of tokenised information a model can use in a given interaction, depending on its technical limits.

Continued training of a pre-existing model on new data or objectives to modify its behaviour.

Production by a model of an unsupported or incorrect statement or content presented in a potentially plausible way.

Setting chosen to define a learning or architecture rather than learned directly as a standard weight.

Family of models producing synthetic content such as text, image, audio, video or code.

Execution of models directly on a locally controlled device, without needing to send all data to a remote service.

Use of an already prepared model to produce a result from new inputs.

Field studying and developing systems performing tasks associated with abilities such as perceiving, predicting, reasoning or generating.

Model trained on large text or related corpora to process and produce linguistic sequences.

Parametrised system learned or designed to transform inputs into predictions, classifications, decisions or content.

Model that processes or produces several modalities of information, for example text, image or sound.

Internal value determining a model's behaviour, generally adjusted during training.

Content passed to a model to guide its response, possibly including instructions, questions and documents.

Representation of numeric values with less precision to reduce memory or computation, possibly at the cost of quality.

Architecture combining retrieval of relevant documents and response generation to provide the model with external information in context.

Model composed of interconnected compute units whose parameters are adjusted to represent relationships in data.

Situation where a model that is too constrained or insufficiently trained fails to capture useful regularities in the data.

Situation where a model learns the specificities of training examples too well and generalises less well.

Splitting or conversion of an input into units usable by a model or linguistic analysis system.

Unit of splitting of text or another modality used by a model; it does not necessarily equate to a word.

Neural architecture based notably on attention mechanisms, widely used in modern language models.

Process aiming to make identification of a person impossible in practice according to a re-identification risk analysis.

Legal ground allowing a personal data processing, for example consent, contract or legal obligation depending on the case.

Protection where only the intended endpoints normally possess the means to read the exchanged content, depending on implementation.

Expression of will meeting the applicable legal conditions; it is not the only possible legal basis for processing.

Any information relating to an identified or identifiable natural person, directly or indirectly.

Right to request deletion of personal data in the situations provided by regulation, with exceptions.

Right, under conditions, to obtain confirmation of processing and communication of one's personal data.

Period during which data is kept with regard to the purpose and applicable obligations.

Inference of a probable identifier from characteristics of the browser and device, sometimes without cookies.

Determined objective for which personal data is collected or used.

Principle of processing only data that is adequate, relevant and necessary for the pursued purposes.

Right, in the cases provided, to retrieve certain data provided in a structured format and transmit it elsewhere.

Default application of settings that limit personal data processing to what is necessary for the considered purpose.

Integration of data protection from the design of systems and throughout their operation.

Processing that makes it difficult to attribute information to a person without additional data; data remains personal.

Actor determining, alone or jointly, the purposes and means of a personal data processing.

European regulation governing personal data processing and defining rights and obligations according to their conditions of application.

Organisation processing personal data on behalf of a controller, within instructions and defined obligations.

Collection and matching of behavioural information for targeting or measurement purposes, subject to applicable rules.

Operation on personal data, such as collection, storage, consultation, modification or deletion.

Method of searching for the factors behind an incident to avoid fixes limited to symptoms.

Organised collection of procedures and solutions to capitalise on answers to recurring problems.

Design or implementation defect causing behaviour not matching the intended expectations.

Standardised or program-specific identifier indicating a precise category of failure or state.

Ability of components to function together regarding interfaces, firmware, dimensions and electrical constraints.

Incompatibility between drivers or components causing malfunctions, instability or loss of features.

Undesired alteration of the content or structure of data following hardware, software or human errors.

Methodical approach of identifying the likely cause of an incident, applying a fix and verifying the result.

Analysis of symptoms, logs and tests intended to determine the origin of a malfunction.

Display signalling a critical Windows stop, generally linked to a system error or low-level component.

Symptom of absent display that may relate to power, video, firmware or software boot.

HTTP response code indicating that the server did not find the requested resource.

HTTP code signalling an unexpected internal server error while processing a request.

Failure caused by invalid or unauthorised memory access in certain runtime environments.

Situation in which software or a system temporarily or persistently no longer responds to actions.

Critical halt of certain Unix-like systems when the kernel detects a condition it cannot safely handle.

Replacement of a device's firmware, an operation requiring compatibility and precautions against interruption.

Boot mode loading a reduced set of components to isolate certain causes of failure.

Unexpected termination of a program or system, sometimes accompanied by an error report.

Attempt to restore lost or inaccessible files from backups, metadata or a damaged medium.

Shutdown followed by a new initialisation of software or hardware components, able to resolve some transient states.

Return of settings or a system to a defined state, with possible data loss depending on the chosen operation.

Installing software or a system again to replace missing, corrupted or misconfigured components.

Area of a storage medium that can no longer be read or written reliably according to the device.

Monitoring technology providing certain state indicators for storage media, without predicting all failures.

Value computed over data to detect certain accidental modifications, without always guaranteeing cryptographic authentication.

Examination intended to detect RAM errors using read and write patterns; a negative result does not exclude all failures.

Structured record of a malfunction, its state, priority and actions taken.

Comparison of data or structures to a reference or to consistency rules to detect alterations.

Professional responsible for the operation, configuration and monitoring of network infrastructures.

Professional ensuring configuration, maintenance and operational security of computer systems.

Professional defining a system's major structures and technical choices and their consequences.

Ordered list of work to be done on a product or project, with a varying level of detail.

Person coordinating objectives, resources, deadlines, risks and communication of a digital project.

Professional preparing and analysing data to produce indicators and inform business questions.

Professional using statistics, programming and modelling methods to explore data and produce analyses or predictions.

Professional designing, writing, testing and maintaining code that meets defined needs.

Function or department steering information systems, their operation and their evolution within an organisation.

Service centralising assistance requests, qualifying them and routing them to the right people.

Professional designing the pipelines that collect, transform and make data available.

Management method that visualises workflow and limits work in progress to improve task circulation.

Set of approaches favouring iterations, collaboration and adaptation to feedback, without a single mandatory model.

First version gathering enough features to test a value proposition against real uses and gather feedback.

Limited implementation intended to verify the feasibility of a technical hypothesis, without necessarily being a deliverable product.

Role responsible for maximising a product's value and managing its backlog within the Scrum framework.

Exploratory version intended to test an idea, interaction or solution before full development.

Evolving representation of the objectives, priorities and stages envisaged for a product or programme.

Agile framework based on defined responsibilities, events and artefacts to organise development iterations.

Description of a system's expected behaviours and functions from the needs perspective.

Description of a solution's constraints, architectural choices and implementation modalities.

Fixed-duration work period in Scrum at the end of which a usable product increment is sought.

Coordinated set of human, technical, data and process resources supporting an organisation's activities.

Estimate of the direct and indirect costs of a solution over its useful life, beyond its purchase price alone.

Professional helping users resolve incidents and use equipment or software.

Informal term designating old software that is little or no longer commercialised; this status does not automatically remove copyright.

Adaptation of equipment and peripherals to the diverse capabilities and needs of users, complementing software accessibility.

Shared ledger structured in cryptographically linked blocks according to consensus rules, without by itself guaranteeing the truth of entered data.

Implementation of intensive compute resources and methods for simulations, analyses or large scientific tasks.

Simultaneous execution of parts of a workload across several compute units to save time or handle a load.

Processing performed close to data sources or users rather than only in a remote datacentre.

Software or hardware reproduction of another environment's behaviour to run programs intended for it.

Inequalities in access to equipment, connectivity, skills or digital services.

Organisation of computation or data across several machines communicating to provide a common result.

Computer systems integrated into devices performing a specific function, such as a vehicle, sensor or industrial machine.

Interconnection of physical objects equipped with sensors, actuators or communication means.

Old system kept for its usefulness or dependencies, whose maintenance and evolution may be constrained.

High-capacity system designed to handle many transactions and critical loads with strong availability.

Computer historically built around a microprocessor, at the origin of the rise of personal computing.

Situation where software or a device becomes less usable due to evolving requirements, compatibility or support.

Computer intended for use by one person; commercial usage may also specifically target IBM PC-compatible machines.

Restoration and verification of a previously used device before its remarketing or reuse.

Ability of a device to be diagnosed, disassembled and repaired with appropriate parts, tools and information.

Ability of a recent system to execute or interpret resources designed for a previous generation.

Practice of preserving, collecting, restoring and studying computers, software and uses of past generations.

Approach aiming to limit the environmental impacts of IT by acting on equipment, uses and lifecycle.

Computer intended for demanding professional tasks such as design, scientific computing or graphics.

System gathering very large compute capacities for massive, scientific or industrial parallel problems.

System that must meet explicitly defined temporal constraints, distinct from a mere impression of speed.