The IP frame is called the datagram and there exist two main versions of the IP: IPv4 and IPv6.
The IP header has a minimum size of 20 bytes (if no options are used) and a maximum size of 64 bytes (including options and padding bits). Due to a set of different options that can be appended to the IPv4 header, these headers can become very large.
- Version: IP protocol version, here IPv4.
- Internet header length (IHL length): The length of the header.
- Type of Service: The QoS parameters for IP.
- Total Length: The length of the IP frame including header and payload field.
- Identification, Fragment Offset: Both used in case of fragmentation/reassembly.
- Time to Live: A hop counter to prevent circular routing.
- Protocol: Indicates the higher layer protocol that uses IP as the transport layer; typical examples are ICMP, TCP, UDP.
- Source Address: IP address of the sender of the datagram.
- Destination Address: IP address of the receiver of the datagram.
- Options: For example, the timestamp of each router that the IP packet passed.
- Padding: Fill bits to align the header to a multiple of 32 bits.
Since the maximum packet size of an IP datagram can vary from one local network to the next, the IP is equipped with fragmentation/reassembly functionality that allows the transmission of larger frames in series of smaller portions. Figure 2 shows an example where a frame with 1600 bytes of data is fragmented into two smaller frames with 1480 and 120 bytes of data each. Fragmented frames do all have the same frame ID (in the example: 1234). As long as more fragments are following the first one, the fragmentation flag MF is set to "1." The last frame in a series of fragments has fragmentation flag MF = "0," but a fragmentation offset that is required for proper reassembly on the receiver side.
IP fragmentation (Figure 2) may be found in the user plane data streams, but should be avoided on interfaces that carry 3GPP signaling.
IPv4 addresses are typically written in the so-called dotted decimal notation, for example, 126.96.36.199. There are 32 bits (= 4 bytes) reserved for the address fields in the IP datagram. Each number in the dotted decimal format represents the decimal value of a single byte. The dot "." is used as the separator between the different bytes of the IP address. Figure 3 shows a sample address in binary, hexadecimal, and decimal dotted notation format.
- A larger number of possible address values become available. In IPv4 the number of addresses is limited to 32 bits, which means in turn that 232 (≈4.3 billion = 4.3 × 109) possible values can be addressed. IPv6 provides space for 2128 (=3.4 × 1038) possible address values. This is an improvement by a factor of 296 and reached by a restructuring of the IP header. In the IPv6 header shown in Figure 4, 128 bits (16 bytes) is reserved for source and destination addresses. The larger address ranges available for IPv6 will also allow more direct end-to-end packet routing and, hence, less address translation in network nodes is required and the packet routing in the overall network is expected to be faster and more efficient.
- The automatic configuration of dynamically assigned IP addresses is improved and in turn legacy procedures like DHCP (Dynamic Host Configuration Protocol) become unnecessary.
- IPv6 supports Mobile IP, simplifies renumbering (change of dynamically assigned IP addresses), and allows multihoming of subscribers. The purpose of multihoming is to increase the reliability of Internet connections by using two different Internet service providers simultaneously. If the access to one of the providers is interrupted a redirection of packets via the second connection is possible. Mobile IP means that the subscriber always gets the same IP address assigned, no matter if working at home or traveling around.
- IPsec is integrated into IPv6 to achieve a higher security of IP data transmission, while back in IPv4 no security functions were provided at all.
- All in all, the basic header of IPv6 has a simpler structure compared to the header of IPv4. Although the overall header size is larger than in IPv4 (40 bytes, most of them occupied by the longer IP addresses), there are less basic header fields.
- For the version, the decimal number 6 is encoded as binary bit sequence "0110."
- The IPv6 traffic class indicates the packet priority and should not be mistaken for the traffic class QoS element introduced in 3GPP standards that classifies the throughput sensitivity and delay sensitivity of application services. IPv6 traffic class priority values subdivide into two ranges: traffic, where the source provides congestion control, and non-congestion control traffic.
- The flow label is used for QoS management and encoded in 20 bits. Packets having the same flow label value will be treated with the same priority and reliability. This is important for the routing of packets that contain real-time service data.
- The payload length indicates the size of the payload in octets and is encoded in 16 bits. When cleared to zero, the option is a "Jumbo Payload" (hop by hop). The size of the basic header is not counted by the payload length, but the optional header extensions are included. So payload length + 40 bytes (of basic header) = total length of the IPv6 packet.
- The next header information element specifies the next upper layer protocol of the transported payload such as UDP and TCP. The values are compatible with those specified for the IPv4 protocol field (8 bits). The next header information can also point to optional extension headers. In this case the upper layer payload protocol is not indicated by this field.
- The hop limit field (8 bits) indicates the maximum number of routers that are allowed to be involved in routing an IPv6 packet. It replaces the time to live field of IPv4. If the hop limit reaches the value "zero" the packet will be discarded by the router.
- Source and destination addresses, 128 bits each, represent the sender and receiver of the IPv6 datagram.
IPv6 addresses are normally written as eight groups of 16 bits, where each group is separated by a colon (:). For example, 2001:0db8:85a3:0000:0000:8a2e:0370:7334 is a valid IPv6 address.
To shorten the writing and presentation of addresses, several simplifications to the notation are permitted. Any leading zeros in a group may be omitted; thus, the given example becomes: 2001:db8:85a3:0:0:8a2e:370:7334.
Also, one or any number of consecutive groups of value 0 may be replaced with two colons (::): 2001:db8:85a3::8a2e:370:7334.
It is possible to use IPv6 addresses in the URL notation format. In this case the IPv6 address information is enclosed in square brackets:
The brackets prevent that part of the IPv6 address being misinterpreted as port number information. A URL including IPv6 address and port number looks like this: