Plain-language glossary
These explanations cover the language used across the site and the practical reason each term matters. They are precise enough to support the articles, but not written as though a networking certification exam is hiding at the bottom.
A technical term is useful when it names a real distinction. It is less useful when it replaces an ordinary sentence and leaves the reader to conduct a second research project before finishing the first paragraph.
A collection of computer equipment and software run at home for learning, personal services or both. It can be one small computer or several separate systems. A rack is not mandatory, although racks have excellent lobbying and create the impression that somebody planned all of this.
Why have one? It lets you learn on equipment you control and run useful services at home. A single-purpose appliance is simpler when you only need one job done.
A computer or program that provides something to other devices. A media server provides a library and playback streams; a storage server provides files. The word describes the job, not necessarily an enormous machine in a refrigerated room.
Choose a dedicated server when several devices need the same library or service. For a single person with one device, running the program directly on that device may be easier.
Network-attached storage: a dedicated device that stores files and makes them available over the network. It usually combines several drives with management software for shared folders, health checks and storage protection. It is still one physical device and should not be mistaken for a complete backup strategy.
Choose a NAS when multiple services need shared files or you want storage managed independently. Drives inside one server remove the network dependency and may cost less.
A common method for letting one computer use folders stored on another computer. On this site, the main server uses NFS to reach shared storage on the NAS. If that network path stops answering, programs waiting for those folders may look frozen even when their own code is fine.
Choose network file sharing when several applications need the same folders. Local storage is simpler if no other machine needs those files; either way, know what will wait if the file source goes away.
A software-based computer running inside another computer. It has its own operating system and behaves much like a separate physical machine. This provides strong separation, but it also uses more resources than a lightweight container.
A virtual machine is handy when a program needs a separate operating system or a strong maintenance boundary. A container usually costs less overhead when it only needs to package one application.
A lightweight way to run a program with the files and settings it needs. Containers share more of the host system than virtual machines, so they usually start quickly and use fewer resources. They still need careful storage, network and backup decisions; the word 'container' does not place the consequences in a sealed jar.
A container suits small services that share the host's operating system. Use a virtual machine when a separate operating system or a stronger boundary is worth the extra resources.
A widely used system for building and running application containers. A Docker setup can describe the application, its network access, stored data and relationships with other services. This makes the setup repeatable, provided the description itself remains understandable.
It helps make repeated deployments consistent. For a single uncomplicated program, a normal installation can be easier to understand and maintain.
A way for a container to keep important data outside its temporary internal filesystem. The application can be replaced or updated while its settings and database remain. If the location is missing from the backup plan, the container may return with the memory of a goldfish.
Use one for data you need after an application update. Back up that data separately; replacing the container should not replace your only copy of its settings.
The device that moves traffic between the home network and other networks, including the internet. It commonly handles firewall rules and address assignment. Some gateways can also report or block suspicious traffic; that feature is not a substitute for keeping the devices behind them updated.
The provider-supplied router may be perfectly adequate. More advanced gear earns its cost when you genuinely need better visibility, coverage placement or network controls.
A device that connects several wired devices on the same local network. A managed switch can also report link speeds, divide traffic into separate networks and power compatible equipment. It does not replace the router; the two devices have different jobs despite both having many sockets.
An unmanaged switch is fine when you merely need more wired sockets. Managed features are useful when you need to see link problems or separate traffic, but someone must maintain the settings.
A device that provides Wi-Fi and connects wireless devices to the wired network. In some homes the router and access point are combined in one box. Separating them allows access points to be placed where coverage is needed rather than where the internet cable happens to enter the building.
A separate access point helps when Wi-Fi coverage is poor but the router itself works well. Moving the wireless signal closer to the users can be more effective than buying a more expensive router in the same corner.
Power over Ethernet: electricity and network data travel through the same cable. It is commonly used for access points, cameras and similar equipment. The switch must have enough total power available for all connected devices, not merely enough sockets.
Power over one cable makes sense for equipment installed far from an outlet. Ordinary power adapters can be cheaper if the devices are already beside sockets.
A cable connection between access points or other network equipment that might otherwise communicate wirelessly. It usually provides a faster and more predictable path and leaves wireless capacity available for actual wireless devices. The cable is less exciting than a mesh diagram and considerably harder to interfere with using a microwave oven.
Use it where a cable is practical and a stable connection matters. Wireless mesh is a reasonable compromise when running a cable is impractical, with performance that depends more on placement and interference.
A protected network connection that sends selected traffic through another server. It can be useful for privacy on an untrusted network or authorized remote access. It is not a guarantee of anonymity; for an application that must always use that route, check what happens if the connection fails.
A VPN changes who carries the traffic, not whether an application is safe or lawful. A whole-device VPN is simpler; app-specific routing avoids changing everything else but needs more careful checks.
A VPN method designed to be relatively small and efficient. A tool such as Gluetun can manage a WireGuard connection for an application. The method itself is separate from the rules that determine which programs must use it.
It is one option among VPN methods and services. Pick a supported setup you can maintain; the firewall and application routes matter just as much as the protocol name.
A design in which access stops when a protection fails. For an application that must use a VPN, that would mean losing internet access if the VPN is unavailable. The opposite, failing open, keeps the app working through a different route. Which behaviour is appropriate depends on the job, but it should be deliberate.
Choose this when using an unprotected route would defeat the purpose of the application. If uninterrupted access matters more, a deliberate fallback may be appropriate, but it should not happen by surprise.
Plex sends the original file to the playback device without changing the video, audio or packaging. This is usually the simplest and least demanding path. It works only when the device and app support the exact combination used by the file.
Prefer it when the playback device handles the file as-is: the server works less and the original quality is retained. If it fails, inspect compatibility before replacing the whole server.
Plex changes the file packaging but leaves the video itself unchanged. This can solve a compatibility problem without the heavier work of converting the video. The name is confusing because Direct Stream is not the same as Direct Play, even though both may leave the video quality intact.
This is a useful middle ground when the video is supported but its file packaging is not. It generally costs less processing than converting the picture itself.
Converting video, audio or both while a file plays because the playback device cannot use the original format. It requires processing power and can reduce quality. The active Plex session shows whether conversion is happening, which is more useful than assuming the server is doing nothing because the television looks expensive.
It helps an incompatible device play a file, at the cost of server work and sometimes picture quality. A compatible playback device or file can avoid that work entirely.
The method used to compress and decode video or audio. Two files with the same resolution may use different codecs and behave differently on the same playback device. Support depends on the device, application and sometimes the exact profile used inside that codec.
Check it when two files at the same resolution behave differently. Choosing a format your regular devices support can be more useful than buying a faster server to convert every file.
The file packaging that holds video, audio, subtitles and related information together. The packaging and the video codec are separate properties. A device may understand the video inside a file but require Plex to move it into different packaging before playback.
Changing only the packaging may fix a playback problem without changing picture quality. That is worth checking before re-encoding an entire video.
The amount of data used over time by video or audio. Higher bitrate can preserve more detail, but it also requires more storage and a connection that can deliver the data consistently. Resolution alone does not reveal bitrate or overall quality.
It helps explain why two files labelled 1080p can differ greatly in size and appearance. Higher is not always better if the network or storage budget cannot support it.
High dynamic range: video information intended to provide a wider range of brightness and colour on compatible displays. Several HDR formats exist, and support varies by television, playback device and app. A 4K label does not automatically guarantee that every HDR version will display correctly.
Use it when the screen and playback chain support the same format. A predictable standard-range version is better than HDR that the display handles incorrectly.
A point-in-time view of supported storage. It can preserve earlier versions efficiently and make recovery from deletion or unwanted changes faster. Because snapshots may still refer to deleted data, they can also explain why free space does not return immediately.
Use snapshots for quick rollback after deletion or a bad change. Keep a separate backup too: a snapshot on a lost device is lost with it.
A separate recoverable copy of important data or a complete service. A backup should survive the failure it is intended to protect against and be tested enough that restoration is understood. Merely possessing a file with 'backup' in its name is an encouraging first draft.
Choose the kind of copy for the failure you want to survive, then practise a safe restore. A green completion icon cannot tell you whether the missing parts are included.
A way of combining several drives to improve availability, performance or both, depending on the layout. It can allow a storage system to continue after a drive fails. It is not a backup because deletion, corruption or loss of the whole device can affect the entire array.
Use it to keep storage available through some drive failures, not as a replacement for another copy. It usually copies accidental deletion just as faithfully as your good data.
A service that receives an outside web connection and passes it to the appropriate internal web application. It can centralize certificates, names and access rules. It adds a useful control point, but it does not automatically make the application behind it secure.
Useful when several web apps need a shared entry point and consistent access rules. For one service, the extra component may add more upkeep than value.
A router rule that directs selected incoming internet traffic to a service inside the network. It is sometimes required for remote access. Every forwarding rule should have a current purpose, because an old experiment can otherwise remain publicly reachable long after everybody forgot why it existed.
Use only if a supported service genuinely needs inbound internet access. Other remote-access methods may reduce direct exposure, but none removes the need for updates and strong accounts.
Intrusion detection watches network traffic for known suspicious patterns and reports matches. Intrusion prevention can also block matching traffic. An alert proves that a rule matched; the recorded action and surrounding evidence determine whether the attempt was merely observed, blocked or associated with something more serious.
These can provide useful evidence and blocks, but false positives and ambiguous warnings are normal. A basic firewall, current software and careful account security still matter.
The system that turns a readable name into the network address needed to reach a service. When DNS fails, the service may still be running but appear unavailable by name. This makes DNS a remarkably small part of the system with an ambitious ability to ruin the afternoon.
Check DNS when a service works by its direct connection but not by its name. It is a naming system, not proof that the destination itself is healthy.
A way to create separate logical networks on managed equipment without needing a different physical switch for each one. It can limit which groups of devices communicate. The separation is useful only when the related routing and firewall rules are understood and maintained.
Useful when distinct groups of devices need different access rules. For a small home network, one well-maintained network can be safer than several poorly understood segments.