# Network topologies
Z-Mesh is topology-agnostic. It is the physical layer that defines the topology. All types are supported, the only requirement is that Content Stores, devices and apps can send and receive packets.
Being a network layer protocol, Z-Mesh is not dependent on any particular type of physical layer. It can run on wired or wireless, broadcast or directed connection technologies. These could be wireless radios (868MHz/915MHz/2.4GHz, any modulation form), GSM/LTE, Ethernet or tunneled in UDP/IP packets. Please see the Physical layer mapping specification for details.
Typical types of network topologies are:
| Centralized | Decentralized | Distributed |
|---|---|---|
| All nodes connect to, and depend on, a single central authority or hub. | Multiple semi‐autonomous hubs (or clusters), each serving a subset of nodes; hubs may interconnect. | No central hub - every node can act independently, both as client and server, sharing responsibility for data, services, and decision-making. |
# Physical layer independence drives the Network Effect
Being independent of the underlying physical layer and topology is key to driving a strong network effect. There are several reasons for this:
Broad Participation: Any device or user can join regardless of how they connect (Wi-Fi, GSM/LTE, radio-mesh, IP), maximizing the total pool of participants—and hence the network’s value.
Seamless Growth: You can add new nodes or technologies without redesigning the core protocols or services. That freedom fuels rapid, organic expansion.
Interoperability: Abstracting away physical-level details lets different vendors, platforms, and regions interconnect smoothly, avoiding "walled gardens" that fragment the user base.
Resilience and Flexibility: If one physical medium or path fails or becomes congested, traffic can reroute over others without disrupting the higher-level service - keeping the network robust as it grows.
Innovation and Competition: Decoupling services from specific hardware encourages third parties to build new applications or access methods (e.g., new radio standards, edge-computing nodes) that all tap into the same thriving ecosystem.
# Wireless technologies
Z-Mesh works on (almost) any type of wireless physical layer, be it mesh, NIDD, directed radio (satelite) or light (LiFi).
# Light
LiFi (Light Fidelity) is a wireless‐networking technology that uses visible (and sometimes infrared or ultraviolet) light to transmit data instead of radio waves. It has an ultra-high throughput and is secure as only the two ends will see the beam. It is also ideal in RF-free- or RF-hostile environments radio is restricted or could cause interference (e.g., hospitals, aircraft cabins, petrochemical plants).
# Mesh
A wireless mesh network can extend the coverage as each node relays traffic. Adding nodes broadens the network’s reach. As there are multiple redundant paths, data can be rerouted automatically around failures or interference, thus improving reliability.
There are a few types of mesh networks, some more controlled that others. To name a few:
- Broadcast Mesh: Every node blindly rebroadcasts incoming packets to all of its neighbors. Simple to implement, redundancy in traffic, but at the cost of high overhead.
- Directed forwarding: Nodes maintain routing information and forward packets only along calculated paths toward their destination. Efficient, scalable but more complex to implement.
- Time slotted: Uses a synchronized time schedule (e.g., TDMA) so nodes transmit in assigned slots, often combined with routing. Effecient, predictable latency but complex as it requires time-synchronization.
# GSM/LTE Non-IP Networking (NIDD)
Cellular Non-IP Data Delivery (often called NIDD) is a "control-plane" IoT transport option in 3GPP cellular networks (LTE-M, NB-IoT and 5G) that carries small messages without ever setting up an IP session. Instead of tunneling packets over a user-plane PDP/PDN connection, the device sends and receives application data encapsulated inside NAS (Non-Access Stratum) messages via the network’s control signaling.
Key benefits of NIDD:
- Minimal overhead—no IP or TCP/UDP headers, so very small payloads (tens of bytes) are space-efficient.
- Lower latency for tiny messages—no need for PDP context activation or IP address assignment.
- Better battery life—avoids the power and signaling cost of user-plane data bearers.
- Simplified core integration—IoT platforms can interface to the operator’s SCEF (Service Capability Exposure Function) or an NIDD APN to exchange data directly with devices.
Because of these efficiencies, NIDD is widely used for low-throughput, delay-tolerant IoT applications such as metering, environmental sensing, asset-tracking and simple command/control.
# Delay-Tolerant Networking
Delay-Tolerant Networking (DTN) enables (delayed) data communication over networks where end-to-end connectivity is intermittent, latency is very high, or links are frequently disrupted. By decoupling the sender and receiver in time and leveraging persistent storage along the path, DTN handles delays from seconds to hours or days.
Z-Mesh, being an Information-Centric Networking architecture, stores the data in the network-nodes even after forwarding. This means that when connectivity is restored, content can then be re-transmitted.