Skip to content

ip.md

Latest commit

 

History

History
237 lines (143 loc) · 13.7 KB

ip.md

File metadata and controls

237 lines (143 loc) · 13.7 KB

English | 中文版

Internet Protocol (IP)

[TOC]

IP is the network‑layer workhorse of the Internet protocol suite. It provides a best‑effort, connectionless datagram service that carries transport protocols (TCP, UDP) and control protocols (ICMP, IGMP). IP does not guarantee delivery, ordering, or duplicate suppression — those are handled by higher layers when needed.

Types

IP addresses can be classified in several ways based on their structure, purpose, and the type of network they are used in.

types_of_ip_addr

Based On Addressing Scope

public_vs_private_ip_addr

Public IP Addresses

A public IP address is assigned to every device that directly accesses the internet. This address is unique across the entire internet.

Private IP Addresses

Private IP addresses are used within private networks and are not routable on the Internet. It enables devices to exchange data internally.

The Three Private Ranges:

Class Range CIDR Number of Addresses Typical Use
Class A 10.0.0.0 – 10.255.255.255 10.0.0.0/8 16,777,216 Large enterprises, cloud VPCs
Class B 172.16.0.0 – 172.31.255.255 172.16.0.0/12 1,048,576 Medium-sized organizations
Class C 192.168.0.0 – 192.168.255.255 192.168.0.0/16 65,536 Home networks, small offices

Based on IP Version

IPv4

ipv4_addr_format

IPv4 is the most common form IP Address. It consists of four sets of numbers(octets) separated by dots. Each octet represents eight bits, or a byte, and can take a value from 0 to 255.

IPv6

ipv6_addr_format

IPv6 addresses were created to deal with the shortage of IPv4 addresses. They use 128 bits instead of 32, offering a vastly greater number of possible addresses. These addresses are expressed as eight groups of four hexadecimal digits, each group representing 16 bits. The groups are separated by colons.

Based on Assignment

Static IP Addresses

Static IP Addresses are permanently assigned to a device, typically important for servers or devices that need a constant address.

Dynamic IP Addresses

Dynamic IP Addresses are temporarily assigned from a pool of available addresses by the Dynamic Host Configuration Protocol (DHCP).

Based on Function

unicast_multicast_broadcast

Unicast Address

In unicast, data is sent from one sender to one specific receiver identified by a unique IP address.

Broadcast Address

In broadcast, a message is sent from one device to all devices in the same network segment. Every device in the network receives and processes the broadcast message. Its purpose is one-to-all communication within a network.

Multicast Address

In multicast, data is sent from one source to multiple selected receivers that are part of a multicast group. Only devices that have joined the group will receive the data, making it more efficient than broadcasting. Its purpose is one-to-many (selected group) communication.

Anycast Address

In anycast, data is sent from one sender to the nearest receiver (in terms of network distance) among a group of devices sharing the same IP address. Routers determine the closest destination dynamically. Its purpose is one-to-nearest communication (based on routing distance).

Special IP Addresses

There are also some special-purpose IP addresses that don't follow the usual structure:

Address Range CIDR Purpose RFC
0.0.0.0/8 0.0.0.0 – 0.255.255.255 "This host on this network" (used during boot, default route) RFC 1122
127.0.0.0/8 127.0.0.0 – 127.255.255.255 Loopback – localhost (127.0.0.1 most common) RFC 1122
169.254.0.0/16 169.254.0.0 – 169.254.255.255 Link-local (APIPA) – when DHCP fails, self-assigned RFC 3927
224.0.0.0/4 224.0.0.0 – 239.255.255.255 Multicast – one-to-many communication RFC 5771
240.0.0.0/4 240.0.0.0 – 255.255.255.254 Reserved (formerly "Class E") – future use RFC 1112
255.255.255.255 Single address Limited broadcast – sends to all on local network RFC 919

IPv4 header

ipv4_head_struct

Key IPv4 header fields:

  • Version (4 bits): protocol version (4 for IPv4).
  • IHL (Internet Header Length, 4 bits): header length in 32‑bit words (min 5, i.e., 20 bytes without options).
  • DSField / TOS (8 bits): Differentiated Services / Type of Service for QoS markings.
  • ECN (2 bits): Explicit Congestion Notification bits.
  • Total Length (16 bits): entire datagram size in bytes (header + data).
  • Identification (16 bits): identifies fragments of the same original datagram.
  • Flags (3 bits): control fragmentation (DF = don't fragment, MF = more fragments).
  • Fragment Offset (13 bits): location of this fragment's data relative to original datagram (in 8‑byte units).
  • TTL (Time‑to‑Live, 8 bits): maximum number of hops (routers) the datagram may traverse; decremented by each router.
  • Protocol (8 bits): indicates the encapsulated transport protocol (e.g., TCP=6, UDP=17, ICMP=1).
  • Header Checksum (16 bits): covers the IPv4 header only; recomputed at each hop if header fields change.
  • Source / Destination IP addresses (32 bits each).
  • Options (variable): rarely used; extend header functionality (security, routing, timestamps).

Notes:

  • The maximum IPv4 datagram size is 65,535 bytes (Total Length field). Larger data must be fragmented at the IP layer or at the sender's transport layer.
  • Fragmentation/reassembly imposes performance and reliability costs; avoid fragmentation when possible (Path MTU Discovery).

Fragmentation and reassembly

  • When an IPv4 datagram exceeds a link's MTU, a router (or sender) may fragment it into smaller datagrams. Each fragment carries the same Identification value and appropriate Fragment Offset and MF flag.
  • The destination reassembles fragments using Identification and offsets. If fragments are lost, the entire original datagram cannot be reassembled and must be retransmitted by transport (if reliable) or lost.
  • The DF (Don't Fragment) flag prevents fragmentation; if set and the MTU is too small, the router drops the packet and sends an ICMP "Fragmentation Needed" message (used by PMTUD).

Internet checksum (IPv4)

IPv4 uses a 16‑bit one's complement checksum computed over the header. The checksum helps detect header corruption; it does not protect the payload. Routers must recompute the checksum if they modify header fields (for example, decrementing TTL).

Example verification (diagram):

ip_verify

Note: IPv6 removes the header checksum to avoid per‑hop recomputation and improve performance.

IPv6 header

ipv6_head_struct

Key IPv6 header fields:

  • Version (4 bits): 6 for IPv6.
  • DSField (8 bits) and ECN: QoS fields similar to IPv4.
  • Flow Label (20 bits): optional label to identify flows for special handling.
  • Payload Length (16 bits): length of payload after the 40‑byte fixed header.
  • Next Header (8 bits): identifies the type of the first header after the IPv6 header (transport protocol or extension header).
  • Hop Limit (8 bits): like IPv4 TTL; decremented by each router.
  • Source / Destination Addresses (128 bits each).

Notable differences from IPv4:

  • Fixed 40‑byte IPv6 header (no header checksum, fewer variable fields), which simplifies and speeds up forwarding.
  • Fragmentation is only performed by the source host (intermediate routers do not fragment in IPv6). Hosts use Path MTU Discovery to avoid fragmentation.
  • Extension headers (Hop-by-Hop, Routing, Fragment, Destination Options, Authentication, Encapsulation) are chained after the main header when needed.

Why IPv6?

IPv6 was designed to address IPv4 limitations: vastly larger address space (128‑bit), simpler header processing, improved support for extension headers, and built‑in features for autoconfiguration and mobility.

Mobile IP (brief)

Mobile IP provides mobility support by allowing a mobile node to receive packets at a care‑of address while maintaining a permanent home address. Mobile IP involves home agents, foreign agents, and tunneling (IP‑in‑IP or other encapsulation). Modern mobility solutions increasingly rely on higher‑layer or network‑based mechanisms, but Mobile IP remains a canonical example in protocol literature.

Summary

Public IP vs Private IP

Private IP Address Public IP Address
Used within a local or private network Used for communication over the internet
Not routable on the public internet Routable on the public internet
Scope is limited to the local network Scope is global
Assigned by router or DHCP server Assigned by ISP
Unique within a local network Globally unique
Requires NAT for internet access Does not require NAT
Hidden from external networks Visible on the internet
Uses reserved private IP ranges Uses globally assigned IP ranges
Freely usable inside networks May involve additional ISP cost

Addressing and basic subnetting

  • IPv4 addresses are 32 bits, commonly expressed in dotted decimal (a.b.c.d). Subnetting uses network prefixes (e.g., /24) to divide address space.
  • IPv6 addresses are 128 bits, expressed in hex colon notation and compressed forms (e.g., 2001:db8::/32). IPv6 strongly encourages subnetting by /64 links for SLAAC.

Practical notes:

  • Use CIDR (classless inter-domain routing) to allocate and route IPv4 prefixes efficiently.
  • For IPv6, prefer stable addressing policies and avoid relying on temporary addresses for long‑term services.

Path MTU Discovery (PMTUD)

  • PMTUD lets endpoints discover the minimum MTU along a path to avoid fragmentation. In IPv4, the sender probes with DF set; routers that cannot forward send ICMP "Fragmentation Needed" messages with the next‑hop MTU.
  • In IPv6, fragmentation must be handled by the source, so PMTUD is essential. ICMPv6 carries the "Packet Too Big" message for this purpose.

IPv4 vs IPv6

IPv4 IPv6
Uses a 32-bit IP address. Uses a 128-bit IP address.
Uses decimal dot-separated notation (e.g., 192.168.0.1). Uses hexadecimal colon-separated notation (e.g., 2001:db8::1).
Provides a limited address space of about 4.3 billion addresses. Provides an extremely large address space for future growth.
Supports manual configuration and DHCP. Supports SLAAC, DHCPv6, and manual configuration.
End-to-end connectivity is often affected due to NAT. End-to-end connectivity is restored without NAT.
IPsec support is optional. IPsec support is built into the protocol design.
Fragmentation is performed by both the sender and the routers. Fragmentation is performed only by the sender.
Does not support flow-based packet identification. Uses a Flow Label field for packet flow identification.
Includes a header checksum. Does not include a header checksum.
Supports broadcast communication. Uses multicast and anycast instead of broadcast.
Header size is variable (20–60 bytes). Header size is fixed at 40 bytes.
Uses address classes (A, B, C, D, and E). Does not use address classes.
Supports Variable Length Subnet Masking (VLSM). Uses prefix-based addressing.

References

[1] James F. Kurose and Keith W. Ross. COMPUTER NETWORKING: A Top-Down Approach. 6th ed.

[2] RFC 791 — Internet Protocol (IPv4)

[3] RFC 8200 — Internet Protocol, Version 6 (IPv6) Specification

[4] What is an IP Address?

[5] Public and Private IP addresses

[6] Reserved IP Addresses

[7] Difference Between IPv4 and IPv6