Cisco Networking: Data Framing
Data travels over the physical media of the Ethernet network in small containers, or frames. There are different methods of framing Ethernet data, but the two that you are likely to see are Ethernet II and IEEE 802.3. The structure of these frames is similar, and the following explains each:
Ethernet II: It is the standard and most often seen framing type and is used to support required headers used by IP.
IEEE 802.3: You should know this frame type is an expanded frame type that Novell chose to use to support its proprietary Internetwork Packet Exchange/Sequenced Packet Exchange (IPX/SPX) protocol. Unless you are on a 1990s Novell network, you are not likely going to encounter the 802.3 Ethernet frame type.
As shown, the standard Ethernet II frame has the following parts:
Preamble: The preamble is a unique sequence of alternating zeros and ones followed by two ones that is eight bytes in length and sets the start of the Ethernet frame. This series of pulses is picked up by the network card in your computer and, by design, this sequence will never occur in the middle of the frame.
Ethernet header: The Ethernet header is made up of the following pieces:
Destination address: Six bytes that contain the MAC address of the NIC that is the target for the network frame.
Source address: Six bytes that contain the MAC address of the NIC that sends the data on physical media.
Type: Two bytes that denote the frame type. The type field identifies the higher layer protocol, which typically is IP.
Data: Between 46–1500 bytes of data. If the data is fewer than 46 bytes, padding is added to bring the frame to the minimum 64 byte frame size. There must be at least 64 bytes of data between preamble sequences.
Frame check sequence (FCS): Four bytes of FCS data is stored at the end of the frame. Prior to sending the frame, the source computer generates a result from the data found in the frame and stores the result in the last four bytes of the frame.
To generate this FCS value, the entire frame is broken into blocks. All those blocks are then added together, and FCS is a sum of all these blocks of data. The receiving computer calculates its own result from the data in the frame and compares the number it calculates to the FCS data. If the results do not match, the frame is considered to be damaged or inaccurate so the frame is discarded.
Some people will also refer to this as cyclic redundancy check (CRC) data or a CRC sum. The purpose of the CRC and FCS is the same, which is to verify that the data that was received was not altered or damaged during transmission.
All network frames have the same basic structure as the Ethernet II frame, regardless of the type of data that they contain.
Even though the frame exists at the physical level, it is often referred to at the data link layer, because the only difference between the data link data and the physical layer structure is the preamble and FCS data, which is data that is not passed from the physical layer to the data link layer. Each layer applies a header to the data that will eventually be sent across the network.
Almost everyone will describe this data structure as the Ethernet frame, and will place it at the physical layer. For the purpose of moving you through the OSI model, you can consider both the physical and data link layers to have been covered.