IEEE 802.2 is the name given to a subsection of the IEEE 802 standard that describes a software component of a computer network. It defines Logical Link Control (LLC), which is the upper portion of the data link layer of the OSI Model. The LLC sublayer presents a uniform interface to the user of the data link service, usually the network layer. Beneath the LLC sublayer is the Media Access Control (MAC) sublayer, which is dependent on the particular medium being used (Ethernet, token ring, FDDI, 802.11, etc.).
The IEEE 802.2 sublayer adds some control information to the message created by the upper layer (typically, the network layer) and passed to the LLC for transmission to the next hop node. The resulting packet is generally referred to as LLC Protocol Data Unit (PDU) and the additional field added by the LLC sublayer for the so-called LLC HEADER. The LLC Header consist of three fields, namely DSAP (Destination Service Access Point) and SSAP (Source Service Access Point), Control. S-SAP is an 8-bit long field that represents the logical address of the network layer entity that has created the message. D-SAP is an 8-bit long field that represents the logical addresses of the network layer entity intended to receive the message. Control field may be either 8 or 16 bit long and, besides defining the packet format, its used to carry some further control information for auxiliary services, such as flow control. Despite the SAP fields being 8-bit long, some bits have specific significance, so that there is room for only 64 distinguished SAP numbers, which are globally assigned by the IEEE to uniquely identify well established international standards. IP does not have an assigned SAP number, because only “international standards” could be given globally assigned SAP numbers. Protocols which are not international standards can use a SAP number from the locally administered SAP number space. The Subnetwork Access Protocol (SNAP) allows EtherType values to be used to specify the protocol being transported atop IEEE 802.2, and also allows vendors to define their own protocol value spaces.
IEEE 802.2 provides two connectionless and one connection-oriented operational modes:
- Type 1 is an unacknowledged connectionless mode for a datagram service. It allows for sending frames
The use of multicasts and broadcasts reduce network traffic when the same information needs to be propagated to all stations of the network. However the Type 1 service provides no guarantees regarding the order of the received frames compared to the order in which they have been sent; the sender does not even get an acknowledgment that the frames have been received.
- Type 2 is a connection-oriented operational mode. Sequence numbering ensures that the frames received are guaranteed to be in the order they have been sent, and no frames are lost.
- Type 3 is an acknowledged connectionless service. It supports point-to-point communication only .
The 802.2 header includes two eight-bit address fields, called service access points or SAPs in OSI terminology; they are the destination SAP (DSAP), and the source SAP (SSAP). The low-order bit of the DSAP indicates whether it contains an individual or a group address. If the low-order bit is 0, the remaining 7 bits of the DSAP specify an individual address, which refers to a single local service access point (LSAP) to which the packet should be delivered. If the low-order bit is 1, the remaining 7 bits of the DSAP specify a group address, which refers to a group of LSAPs to which the packet should be delivered. The low-order bit of the SSAP indicates whether the packet is a command or response packet; if it's 0, the packet is a command packet, and if it's 1, the packet is a response packet. The remaining 7 bits of the SSAP specify the LSAP from which the packet was transmitted.
Some protocols, or families of protocols, have one or more SAPs assigned to them; for example, IPv4 has a SAP value of hex 06. Those protocols operate directly on top of 802.2 LLC, which provides both datagram and connection-oriented network services.
The Subnetwork Access Protocol (SNAP) can be used with IEEE 802.2; it allows EtherType values to be used with all IEEE 802 protocols, as well as supporting private protocol ID spaces. When both the DSAP and the SSAP are set to the hexadecimal value 0xAA (or 0xAB, if the low-order bit of the field is set), the SNAP service is requested.
Most IEEE 802 protocols, as well as FDDI, always use 802.2. IEEE 802.3 is an exception; Ethernet originally used Ethernet II framing, with a 2-octet type field containing an EtherType value, and no 802.2 header, but the original IEEE 802.3 specification used a framing wherein the 2-octet field following the 6-octet destination and source MAC addresses contained a frame length, with an 802.2 LLC header following the length field. In IEEE 802.3x-1997, the IEEE Ethernet standard was changed to explicitly allow the use of the 16-bit field after the MAC addresses to be used as a length field or a type field.
There exists an Internet standard, RFC 1042, for encapsulating IP version 4 traffic in IEEE 802.2 frames with LLC/SNAP headers. It is almost never implemented on Ethernet, although it is used on FDDI and on token ring, IEEE 802.11, and other IEEE 802 networks.
IP traffic can not be encapsulated in IEEE 802.2 LLC frames without SNAP because, although there is an LLC protocol type for IP, there is no LLC protocol type for ARP. IP Version 6 can also be transmitted over Ethernet using IEEE 802.2 with LLC/SNAP, but, again, that's almost never used (although LLC/SNAP encapsulation of IPv6 is used on IEEE 802 networks).
- Unnumbered format PDUs, or U-format PDUs, with an 8-bit control field, which are intended for connectionless applications;
- Information transfer format PDUs, or I-format PDUs, with a 16-bit control and sequence numbering field, which are intended to be used in connection-oriented applications;
- Supervisory format PDUs, or S-format PDUs, with a 16-bit control field, which are intended to be used for supervisory functions at the LLC (Logical Link Control) layer.
Of these three formats, only the U-format is commonly used. The format of a PDU frame is identified by the lower two bits of the first byte of the control field.