The networking IP stack is modular and highly configurable via build-time configuration options. You can minimize system memory consumption by enabling only those network features required by your application. Almost all features can be disabled if not needed.
IPv6 The support for IPv6 is enabled by default. Various IPv6 sub-options can be enabled or disabled depending on networking needs.
Developer can set the number of unicast and multicast IPv6 addresses that are active at the same time.
The IPv6 address for the device can be set either statically or dynamically using SLAAC (Stateless Address Auto Configuration) (RFC 4862).
The system also supports multiple IPv6 prefixes and the maximum IPv6 prefix count can be configured at build time.
The IPv6 neighbor cache can be disabled if not needed, and its size can be configured at build time.
The IPv6 neighbor discovery support (RFC 4861) is enabled by default.
Multicast Listener Discovery v2 support (RFC 3810) is enabled by default.
IPv4 The legacy IPv4 is supported by the networking stack. It cannot be used by IEEE 802.15.4 or Bluetooth IPSP as those network technologies support only IPv6. IPv4 can be used in Ethernet based networks. By default IPv4 support is disabled.
DHCP (Dynamic Host Configuration Protocol) client is supported (RFC 2131).
The IPv4 address can also be configured manually. Static IPv4 addresses are supported by default.
Dual stack support. The networking stack allows a developer to configure the system to use both IPv6 and IPv4 at the same time.
UDP User Datagram Protocol (RFC 768) is supported. The developer can send UDP datagrams (client side support) or create a listener to receive UDP packets destined to certain port (server side support).
TCP Transmission Control Protocol (RFC 793) is supported. Both server and client roles can be used the application. The amount of TCP sockets that are available to applications can be configured at build time.
BSD Sockets API Support for a subset of a BSD sockets compatible API is implemented. Both blocking and non-blocking datagram (UDP) and stream (TCP) sockets are supported.
Secure Sockets API Experimental support for TLS/DTLS secure protocols and configuration options for sockets API. Secure functions for the implementation are provided by mbedTLS library.
MQTT Message Queue Telemetry Transport (ISO/IEC PRF 20922) is supported. A sample MQTT publisher client application for MQTT v3.1.1 is implemented.
LWM2M OMA Lightweight Machine-to-Machine Protocol (LwM2M specification 1.0.2) is supported via the “Bootstrap”, “Client Registration”, “Device Management & Service Enablement” and “Information Reporting” interfaces. The required core LwM2M objects are implemented as well as several IPSO Smart Objects. (LwM2M specification 1.1.1) is supported in similar manner when enabled with a Kconfig option. LwM2M client sample implements the library as an example.
DNS Domain Name Service (RFC 1035) client functionality is supported. Applications can use the DNS API to query domain name information or IP addresses from the DNS server. Both IPv4 (A) and IPv6 (AAAA) records can be queried. Both multicast DNS (mDNS) (RFC 6762) and link-local multicast name resolution (LLMNR) (RFC 4795) are supported.
Network Management API. Applications can use network management API to listen management events generated by core stack when for example IP address is added to the device, or network interface is coming up etc.
Wi-Fi Management API. Applications can use Wi-Fi management API to manage the interface, in example to connect to Wi-Fi network and to scan available Wi-Fi networks.
Wi-Fi Network Manager API. Wi-Fi Network Managers can now register themselves to the Wi-Fi stack. The Network Managers can then implement the Wi-Fi Management API and manage the Wi-Fi interface.
Multiple Network Technologies. The Zephyr OS can be configured to support multiple network technologies at the same time simply by enabling them in Kconfig: for example, Ethernet, Wi-Fi and 802.15.4 support. Note that no automatic IP routing functionality is provided between these technologies. Applications can send data according to their needs to desired network interface.
Minimal Copy Network Buffer Management. It is possible to have minimal copy network data path. This means that the system tries to avoid copying application data when it is sent to the network.
Virtual LAN support. Virtual LANs (VLANs) allow partitioning of physical ethernet networks into logical networks. See VLAN support for more details.
Network traffic classification. The sent and received network packets can be prioritized depending on application needs. See traffic classification for more details.
Time Sensitive Networking. The gPTP (generalized Precision Time Protocol) is supported. See gPTP support for more details.
Network shell. The network shell provides helpers for figuring out network status, enabling/disabling features, and issuing commands like ping or DNS resolving. The net-shell is useful when developing network software. See network shell for more details.
Additionally these network technologies (link layers) are supported in Zephyr OS v1.7 and later:
SLIP (IP over serial line). Used for testing with QEMU. It provides ethernet interface to host system (like Linux) and test applications can be run in Linux host and send network data to Zephyr OS device.
The networking stack source code tree is organized as follows:
This is where the IP stack code is located.
This is where the IP stack layer 2 code is located. This includes generic support for Bluetooth IPSP adaptation, Ethernet, IEEE 802.15.4 and Wi-Fi.
Application-level protocols (DNS, MQTT, etc.) and additional stack components (BSD Sockets, etc.).
Public API header files. These are the header files applications need to include to use IP networking functionality.
Sample networking code. This is a good reference to get started with network application development.
Test applications. These applications are used to verify the functionality of the IP stack, but are not the best source for sample code (see