1. Ethernet frame format
  2. MAC-in-MAC
  3. Asynchronous transmission module

Format of an Ethernet frame

Lead code

An Ethernet frame is preceded by a section called a Preamble, which consists of alternating zeros and ones to indicate the beginning of an Ethernet frame and a signal that the peer network card can ensure synchronization with it. The last two bits of the leading code are a field called SFD[1], which has a value of 11. The leading code and SFD together account for 8 bytes.

The frame body type

Protocol type

Common protocol types are numbered as follows

Type number, agreement
0800 IPv4
0806 ARP
0835 RARP
8100 VLAN
814C SNMP over Ethernet
86DD IPv6
8847-8848. MPLS

MAC-in-MAC

Mac-in-mac, also known as Provider Backbone Bridge (PBB), complies with IEEE 802.1AH. It is a stack Bridge technology that uses MAC headers to support Backbone network traffic engineering and encapsulates user packets into CARRIER MAC frames. In the cloud computing environment, the GENERATION of MAC addresses of all VMS can be controlled. Therefore, when tenant networks in cloud computing are isolated, mac-in-Mac solution, which is an Overlay solution, can be adopted.

The PBB packet encapsulation format encapsulates two MAC header layers. Therefore, a packet usually has two VLAN tags. The VLAN Tag following the outer MAC address header is ETag or BTag, which corresponds to the MAC address of the backbone network and the bridge VLAN ID. The Ethernet Type is 0x880A. The Ethernet Type following the outer Tag is 0x8807 ITag.

In this mode, the MAC addresses of users and carriers are isolated from each other, avoiding overlapping MAC addresses and VLAN ids. In addition, the carrier can control the forwarding of network data packets without affecting the stability of the carrier backbone network and reduce broadcast frames to effectively utilize bandwidth for users in transparent transmission services.

Asynchronous Transfer module (ATM)

ATM is a data link that transmits data in a unit called cell (5Byte header +48Byte data). ATM is mainly used for connection of wide area networks due to its short line occupancy time and efficient transmission of large-capacity data.

Characteristics of ATM

ATM is a connection-oriented data link. Therefore, a communication line must be set before communication transmission. ATM does not have the same send permission restrictions as Ethernet and FDDI, allowing any data to be sent at any time. Therefore, when a large number of hosts send a large amount of data at the same time, it is easy to cause network congestion and even make the network enter the convergence state. In order to prevent this phenomenon, ATM has added a subdivision function to limit the bandwidth.

ATM extends TDM (Time division multiplexing) and can effectively improve line utilization. The sender adds 5-byte information to the packet header, including VPI and VCI codes to identify the specific communication type.

ATM and upper-layer protocols

An Ethernet frame can transmit 1500 bytes at most, FDDI can transmit 4352 bytes at most, while ATM can only transmit a fixed 48 bytes of data. If there is no IP header and TCP header in each cell, upper-layer data cannot be stored. For this purpose, ATM alone is generally not used, but the upper AAL (ATM Adapter Layer) is used.

For upper-layer IP networks, each IP packet can be divided into a maximum of 192 cells after the protocol header of each layer is attached. If one of the 192 cells is lost during data transmission, the entire IP packet is damaged. This is by far the biggest drawback of ARM.

ATM switches

  • When the host directly connects to the ATM switch, the host directly processes the ATM message.
  • When the host is indirectly connected to the ATM switch, the router connected to the ATM switch performs the intellectualization of IP packets.

  1. The Start Frame Delimiter ↩ ︎