Skip to main content

Glossary

Blueprint​

A predefined pipeline that a network sets up automatically on its nodes. A blueprint is written with Go template syntax, and NanoPing fills in values such as node IDs, virtual IP addresses and carriers for each node. A network uses three types of blueprint: Relay (deployed to the relay node), Relay client (deployed to the relay node once for every node in the network) and Client (deployed to every node). NanoPing includes default blueprints for each type. See Creating a blueprint.

Carrier​

One path between two nodes that Raft sends traffic over, usually one per network interface (for example one Ethernet cable, or one cellular modem). Raft spreads a stream across all its carriers and repairs packets lost on any of them. See How Raft works.

Carrier priority​

A number you give each carrier to tell Raft which links to prefer. Higher values are preferred: Raft fills the highest-priority carrier first and uses lower-priority ones only as it nears capacity. Without priorities, traffic is shared round-robin. See Bandwidth-aware load distribution.

Cloud hub​

The hub that NanoPing hosts for your organization. Every device joins it when it is activated, so there is nothing to set up. See Hub.

Cloud library​

Cloud storage for pipelines and blueprints, so you can save a pipeline once and add it to any node. See Creating a pipeline.

Component​

A building block of a pipeline, such as UDP Local, Raft or Peer Writer. Components are connected to each other in the diagram editor. See Components.

FEC – Forward Erasure Correction​

Sending extra repair packets alongside the data, so the receiver can rebuild lost packets without waiting for them to be sent again. Raft uses adaptive FEC, which adjusts the amount of repair data to the loss it measures. NanoPing also has fixed-rate FEC components. See FEC algorithms.

Fixed replication​

A Raft repair mode that sends every packet over all valid carriers. It uses more bandwidth, but a packet lost on one carrier still arrives on another.

Fragmentizer​

A component that splits packets too large for the connections into smaller pieces. It must be paired with a Defragmentizer on the receiving side. A Raft stream splits packets itself, so a pipeline with Raft needs neither. See Packet size and MTU.

Hub​

The control plane of a NanoPing setup: an address book that knows where every node is and tells nodes how to connect to each other. Nodes on the same hub can be managed from any of their dashboards. The hub only coordinates; pipeline traffic goes directly between the nodes. A hub is either your organization's cloud hub or a self-hosted hub server. See Hub.

Jitter​

Variation in how long packets take to arrive. High jitter makes video stutter even when no packets are lost.

Jitter buffer​

A short wait before delivering packets, to smooth out jitter. In NanoPing the jitter buffer is part of Raft and is set per stream with jitter_buffer, in milliseconds, between 0 and the stream's timeout. See Jitter buffer integration.

KPI​

A Key Performance Indicator: a value from a running pipeline, such as loss rate or bandwidth, shown on the dashboard. The KPI page shows the KPIs of every node on the hub. See Quick add KPI.

Lifecycle handler​

The part of a pipeline's configuration that runs commands when the pipeline starts or stops, for example to launch GStreamer in the background. It is set in the pipeline's Code tab under lifecycle_handler.

MTU – Maximum Transmission Unit​

The largest packet a network interface can send, usually 1500 bytes on Ethernet and lower on VPNs such as Tailscale. NanoPing's connections send UDP packets of up to 1200 bytes without an MTU set, which fits nearly any interface. See Packet size and MTU.

Network​

A group of nodes that NanoPing connects automatically, using blueprints to generate the pipelines each node needs, with a relay node in the middle. Each node gets an IP address on the network's virtual network. Networks are managed from the Networks section of the dashboard.

Node​

Any installation of NanoPing. A node runs pipelines and serves a dashboard at http://localhost:8081 by default. A node can also run without an HTTP address, and is then managed from the dashboard of another node on its hub.

Pipeline​

The instructions a node uses to carry traffic: a set of connected components. Pipelines are created from a template, from scratch, from a file, by copying, or from the cloud library. See Creating a pipeline.

Raft​

NanoPing's transport protocol. It takes your application's traffic, spreads it over every carrier, repairs lost packets with adaptive FEC, and delivers the stream in order within its timeout. See How Raft works.

Relay​

A node with a public IP address that forwards packets between two nodes that cannot reach each other directly, for example because both are behind NAT. See AWS Relay Setup.

Session​

The handshake that confirms a carrier reaches the right peer. A pipeline's log shows New Session Bound when a session is up, once per carrier.

Stream​

Your application's traffic going into and out of Raft, such as a UDP video stream. A realtime stream has a timeout and an optional jitter buffer.

Template​

A predefined setup, such as UDP Video Streaming or VPN Tunnel, that creates a matching pair of pipelines through a short guided flow. See Creating a pipeline.

Timeout​

The latency budget of a realtime stream, in milliseconds. Packets older than the timeout are dropped, so it must be higher than the real latency of your slowest link. A timeout of 0 turns packet expiry off. To choose a good value for your application, see Time budgets.

TUN – Network TUNnel​

A TUN (network TUNnel) interface is a virtual network device in the OS that operates at Layer 3 (IP layer). It allows user-space applications to handle IP packets as if they were coming from or going to a real network interface. This is commonly used for VPNs, tunneling, and network simulation.

For more details, refer to the Linux Kernel TUN/TAP documentation.