The Medium Access Control (MAC) protocol is used to provide the data link layer of the Ethernet LAN system. The MAC protocol encapsulates a SDU (payload data) by adding a 14 byte header (Protocol Control Information (PCI)) before the data and appending an integrity checksum, The checksum is a 4-byte (32-bit) Cyclic Redundancy Check (CRC) after the data. The entire frame is preceded by a small idle period (the minimum inter-frame gap, 9.6 or 10.4 microsecond (µS)) and a 8 byte preamble (including the start of frame delimiter).
Preamble
The purpose of the idle time before transmission starts is to allow a small time interval for the receiver electronics in each of the nodes to settle after completion of the previous frame. A node starts transmission by sending an 8 byte (64 bit) preamble sequence. This consists of 62 alternating 1's and 0's followed by the pattern 11. Strictly speaking the last byte which finished with the '11' is known as the "Start of Frame Delimiter". When encoded using Manchester encoding, at 10 Mbps, the 62 alternating bits produce a 10 MHz square wave (one complete cycle each bit period).
The purpose of the preamble is to allow time for the receiver in each node to achieve lock of the receiver Digital Phase Lock Loop which is used to synchronise the receive data clock to the transmit data clock. At the point when the first bit of the preamble is received, each receiver may be in an arbitrary state (i.e. have an arbitrary phase for its local clock). During the course of the preamble it learns the correct phase, but in so doing it may miss (or gain) a number of bits. A special pattern (11), is therefore used to mark the last two bits of the preamble. When this is received, the Ethernet receive interface starts collecting the bits into bytes for processing by the MAC layer. It also confirms the polarity of the transition representing a '1' bit to the receiver (as a check in case this has been inverted).
Header
The header consists of three parts:
The MAC address space is unstructured, in comparison to a hierarchical method. In this scheme blocks of addresses are allocated, and usually used on a first-come first-served basis. Addresses do not denote specific positions within the network topology.
The function of the source address is to allow address-learning to configure the filter tables in a bridge/switch. The addresses associated with interface cards are always source addresses, not multicast nor broadcast, so the second hex digit (reading from the left) will always be even, not odd..
The final field in an Ethernet MAC frame is called a Cyclic Redundancy Check (sometimes also known as a Frame Check Sequence). A 32-bit CRC provides error detection in the case where line errors (or transmission collisions in Ethernet) result in corruption of the MAC frame. Any frame with an invalid CRC is discarded by the MAC receiver without further processing. The MAC protocol does not provide any indication that a frame has been discarded due to an invalid CRC.
The link layer CRC therefore protects the frame from corruption while being transmitted over the physical mediuym (cable). A new CRC is added if the packet is forwarded by the router on another Ethernet link. While the packet is being processed by the router the packet data is not protected by the CRC. Router processing errors must be detected by network or transport-layer checksums.
After transmission of each frame, a transmitter must wait for a period of at least 10.4 microseconds (at 10 Mbps) to allow the signal to propagate through the receiver electronics at the destination. This period of time is known as the Inter-Frame Gap (IFG). While every transmitter must wait for this time between sending frames, receivers do not necessarily see a "silent" period of the full period used at the sender, and must be designed to expect at least 9.6 microseconds. The way in which repeaters operate is such that they may reduce the IFG between the frames which they regenerate.
It is important to realise that nearly all serial communications systems transmit the least significant bit of each byte first at the physical layer. Ethernet supports broadcast, unicast, and multicast addresses. The appearance of a multicast address on the cable (in this case an IP multicast address, with group set to the bit pattern 0xxx xxxx xxxx xxxx xxxx xxxx) is therefore as shown below (bits transmitted from left to right):
0 23 IP Multicast Address Group 47
| | <--------------------------->|
1000 0000 0000 0000 0111 1010 xxxx xxx0 xxxx xxxx xxxx xxxx
| |
Multicast Bit 0 = Internet Multicast
1 = Assigned for other uses
However, when the same frame is stored in the memory of a computer, the bits are ordered such that the least significant bit of each byte is stored in the right most position (the bits are transmitted right-to-left within bytes, bytes transmitted left-to-right):
0 23 47
| | |
0000 0001 0000 0000 0101 1110 0xxx xxxx xxxx xxxx xxxx xxxx
| <--------------------------->
Multicast Bit IP Multicast Address Group
Receiver Processing Algorithm
Any frame which is received and which is less than 64 bytes is illegal, and is called a "runt". In most cases, such frames arise from a collision, and while they indicate an illegal reception, they may be observed on correctly functioning networks. A receiver must discard all runt frames.
Any frame which is received and which is greater than the maximum frame size, is called a "giant". In theory, the jabber control circuit in the transceiver should prevent any node from generating such a frame, but certain failures in the physical layer may also give rise to over-sized Ethernet frames. Like runts, giants are discarded by an Ethernet receiver.
Some modern Gigabit Ethernet NICs support frames that are larger than the traditional 1500 bytes specified by the IEEE. This new mode requires support by both ends of the link to support Jumbo Frames. Path MTU Discovery or PLPMTUD is required to utilise this feature, since there is no other way for a router to determine that all systems on the end-to-end path will support these larger sized frames.
Any frame which does not contain an integral number of received bytes (bytes) is also illegal. A receiver has no way of knowing which bits are legal, and how to compute the CRC-32 of the frame. Such frames are therefore also discarded by the Ethernet receiver.
A protocol known as the "Address Resolution Protocol" (arp) is used to identify the MAC source address of remote computers when IP is used over an Ethernet LAN.
An extension to Ethernet, known as IEEE 802.1p allows for frames to carry a tag. The tag value adds an extra level of PCI to the Ethernet frame header. This increases the size of the total MAC frame when the tag is used. A side effect of this is that NICs and network devices designed to support this extension require a modification to the jabber detection circuit.
- Ethernet
- CSMA/CD and Ethernet Transmit Algorithm
- Ethernet Packet Decodes
- IP packet processing
- MAC Address Assignment
- Manchester Encoding
- Network Interface Cards
- Calculations involving Ethernet (and try the problem sheet)