TCP

Ethernet packets can carry IP packets, which in turn can carry TCP packets.

TCP stands for Transmission Control Protocol. TCP is used to send and receive a stream of data on a pre-established connection. The TCP protocol itself takes care of setting up the connection, and ensuring that all transmitted data is received correctly.

TCP is more reliable than UDP, but is more complex to use and requires more RAM. The additional RAM usage comes from, in part, the necessity to retain packets that have already been sent onto the network until they have been correctly acknowledged (in case the packets need to be retransmitted), and the need to assemble data that is split between multiple packets into a reliable stream.

See also TCP Sockets.

Useful links:

• TCP/IP Networking Basics
• FreeRTOS+TCP networking tutorial

Other networking basics pages:

• Ethernet Networking and Addressing
• MTU - Maximum Transmission Unit
• IP Networking
• IP Network Addressing
• ARP - Address Resolution Protocol
• Subnets / Netmasks
• Gateways and Routers
• UDP Networking
• TCP Networking
• MSS - Maximum Segment Size
• Port Numbers
• Sockets
• Binding
• Clients and Servers
• Static IP Addresses
• DHCP - Dynamic Host Control Protocol
• Name Resolution
• DNS - Domain Name Service
• LLMNR - Link-local Multicast Name Resolution
• NBNS - NetBIOS Name Service
• Bytes Order and Endian