Other meanings of Network packet
Computer Networking
A network packet is a formatted unit of data carried across a packet-switched network, containing both control information (header) and user data (payload). Packets enable efficient multiplexing of multiple communications over shared links, forming the basis of the Internet and most modern data networks.
A network packet generally comprises a header and a payload. The header carries metadata such as source and destination addresses, protocol identifiers, flags, and error-detection fields; the payload contains the actual application data. In layered models, packets are encapsulated as they descend the stack: a transport-layer segment is wrapped in a network-layer packet (e.g., IPv4 or IPv6), which is then placed inside a data-link-layer frame (e.g., Ethernet) for transmission on a physical medium.
The Maximum Transmission Unit (MTU) defines the largest packet payload a link can carry; Ethernet's typical MTU is 1,500 bytes. When a packet exceeds the next hop's MTU, it may be fragmented into smaller pieces, each with its own header, and reassembled at the destination. In IPv4, fragmentation is performed by routers; in IPv6, only the source can fragment (via extension headers).
Packets travel across interconnected networks through routers, which examine the destination address in the header to forward each packet independently. This datagram model (as in IP) contrasts with virtual-circuit networks like ATM or MPLS, which establish a fixed path before data flows. The Internet Protocol (IP) defines the addressing and forwarding rules, while transport protocols such as TCP and UDP segment application data into packets (or datagrams) and reassemble them at the receiver.
Packets may arrive out of order, be duplicated, or be dropped due to congestion; higher-layer protocols handle reliability if needed. The Time-to-Live (TTL) field in IPv4 (hop limit in IPv6) prevents packets from looping indefinitely, being decremented at each router and discarded when it reaches zero.
Beyond classic IP forwarding, packets play roles in numerous specialized contexts. Packet injection in wireless networks (e.g., 802.11) can be used for both security testing and eavesdropping; packet analysis tools like Wireshark dissect headers for diagnostics and forensic study. Deep packet inspection (DPI) inspects payload contents for filtering, shaping, or cybersecurity applications, raising privacy concerns. Network Function Virtualization (NFV) and Software-Defined Networking (SDN) enable programmable packet processing in virtual switches.
Niche historical details include the Implicit Protocol of early ARPANET NCP (predecessor to TCP), and the fact that the packet term was coined by Donald Davies in 1965, independent of Paul Baran's earlier
One overlooked nuance is that a packet is not a physical entity but a logical construct: the same packet may be encapsulated differently at each hop (e.g., VLAN tags added or removed). Also, small packets can cause inefficiency due to per-packet overhead; techniques like TCP segmentation offload reduce CPU load by making the NIC generate packets. In satellite and deep-space communications (e.g., NASA's DTN), packets are replaced by bundles, where store-and-forward replaces hop-by-hop forwarding.
Another edge case is jumbo frames (up to 9,000 bytes) on local networks, which require path MTU discovery. Finally, the term packet in telecom also refers to packet switching in telephony, where voice is digitized and grouped, a concept that predates the Internet and influenced modern VoIP.
This entry focuses on the computer networking sense of 'packet'; for other uses, see packet (disambiguation).
Help improve the encyclopedia. Reports go straight to the site manager.