Ethernet Frame Format: Types, Initial Version, and Internal Modification

Depending on the standard, the Ethernet frame format (frame) is structured, contains more or less information fields of network models. By transmitting data, a frame with a size of 64 to 1518 bytes is responsible for the correct formation of rules and the successful transfer of information to the recipient.

The protocol was created in the 1970s by well-known computer manufacturers: DEC, Intel and Xerox. Initially it was called DIX Ethernet, and later - Thick Ethernet (thick) due to the use of coaxial cable. In the late 1980s, it was upgraded to an Ethernet 2 frame format with support for more features and speeds. Around the same time, IEEE was creating standards for Internet networks. During this time, he jumped up the speed: from 10 to 100 Mbit / s, then 1 Gbit / s, and today - 10 Gbit / s in IEEE 802.3AE.

Frame Format Definition

Ethernet is a selection protocol in local area networks representing a group of connected devices and located relatively close to each other in a limited area. There are factors that distinguish local from global Internet networks:

1. Less geographical coverage.
2. Resources for working with high data rates.
3. Large bandwidth from 100 Mbit / s to 10 Gbit / s, which are used in modernized networks.
4. No lease or internet service provider is required to connect the devices.
5. A LAN can be a small project for an office or a remote worker.

There is a bit of history to know what types of Ethernet frames exist. Ethernet was originally designed to work on a long coaxial cable that connected all the computers on the network. This type is called a tire. When one station transmitted data, all the others received it. It was designed on the assumption that everyone will hear broadcasts on a piece of wire that is used to connect. Hence the terms “wired segment” and “broadcast domain”.

A broadcast domain includes all wires and computers that can hear each other when one of the computers transmits. A wire segment is a part of a wire used to connect two devices. Due to the fact that Ethernet networks consist of broadcast domains, it does not have a clock signal, as is often the case with serial connections.

Instead, systems should determine if the wire is being used, and if not, send enough information so that the remote station can correctly synchronize it. This mechanism, combined with the ability to detect other computers trying to access a wired connection, is a formalized protocol called CSMA / CD-based multi-access discovery.

In the OSI model, the frame is at the data link layer and is responsible for error-free transmission and division of the bit stream into blocks. The first version of Ethernet I was based on 16-bit fields. The modern Ethernet frame format was first used in the Ethernet II framework before the Institute of Electrical and Electronics Engineers developed the standard IEEE 802.3 protocol, the first of which in 1983 was 3raw.

LAN components

A local area network is a traditional environment for connecting users and using resources in the form of data, applications, and other functions with input and output devices such as cameras, PCs, televisions, and printers. One of the most important functions in them is to ensure communication with others. Through default gateways, as well as through routers and WAN peripherals. LAN serves as an entry point to the global network, which provides many resources and the ability to connect to partners, suppliers in Ethernet frame format.

Today, local networks can serve different consumer schemes, from a small office with a pair of devices connected to the Internet, to schemes that combine several buildings and thousands of users. A group of remote employees and mobile users are seen as expanding the corporate network through the use of virtual private VPN lines. The scheme will provide transparent access to a scenario in which user interaction and its operation will be very similar to a direct connection to an enterprise’s local network.

Typical components include endpoints, PCs, servers, IP phones, video consoles, and network devices that provide a unified place to connect. For example, a switch, routers for connecting different networks into one structure, and sometimes more traditional centers as common media. Network cards and cables are also part of this design.

Ethernet frame structure

In this format, all machines can determine the channel and establish the presence of signals from other transmitters. This allows them to detect collisions, which makes the frame part of the discovery protocol. PCs at this time will schedule the transfer according to a random timer, which is different for each machine.

This creates a functional environment that provides good performance in the long run. Some models may be functional, but with poor design, for example, a large collision domain with too many machines using the same channel. This will increase the likelihood of simultaneous transmission and possible collisions, which, in turn, will lead to an overall decrease in performance. Other problems are related to faulty equipment that sends error frames to the network. Therefore, the next important function of any layer 2 protocol is framing.

An Ethernet frame is a container carrying bits that are transmitted on a network that has a field format. For example, Ethernet 2 and IEEE 802.3 represent a sequence of bits used to synchronize two communicating devices and a verification sequence used for integrity, as well as the destination and source addresses — the MAC address.

Communication within the local network

Another important concept in Ethernet and LAN communication is the amount of specific transmission. In unicast there is only one destination - the recipient address representing one machine. This is the most common approach, and the MAC address is a unique identifier that is used to send such frames. If the protocols and applications will send them to all devices on the network, then broadcasting is used.

Broadcast - the destination that is processed by all devices. This is convenient for protocols such as ARP, which request IP address translation to MAC without IP owner information.

Multicast is located in the middle between unicast and broadcast transmissions. This is not one and not all destinations - it is just a specific group. In other words, if the machines belong to a group, then a particular transmission or package will be intended for it. This is more dynamic because machines join and leave groups quickly and efficiently. Example multicast application: video conferencing, e-learning, and other multimedia with a wide reach of users.

MAC addresses are associated with the hardware vendor. IEEE defines ranges to ensure uniqueness that permits their modification for specific purposes. It consists of a 24-bit unique organization identifier or OUI that identifies the equipment manufacturer, whether it is a network card or router ports. These 24 bits have 2 bits with a special meaning:

1. Indication of whether there are broadcast or multicast connections. The managed address bit is typically used when changing the MAC.
2. The 24-bit station address, which is regulated and administered by a particular equipment supplier.

LAN Standards

Ethernet in terms of OSI refers to the data link layer and has specifications at the physical layer. For example, 1 GB and 10 Gbit Ethernet will have physical layer specifications and links to fiber-optic technologies and connectors for higher speeds. The protocol is divided into 2 sublevels: media access control and determining the MAC address as a form of identification for devices on an Ethernet network. The sublevel controlling the logical channel will interact with the upper layers, when the components process each packet, they will refer to the CSMA / CD.

Ethernet 2 frame formats provide MAC addressing and an access method called a CSMA / CD or collision detection multiple access system. This is a point method that, thanks to its unique technology, works for Ethernet and allows all machines to transmit signals at the same time, at any time without priority, since everyone has equal access and is part of the multiple access protocol.

Classic VLAN Model

An Ethernet II frame, also known as V2, is the most common on the Internet. Currently, most TCP / IP applications, such as HTTP, FTP, SMTP, and POP3, carry packets in the following way.

The physical layer format is an Ethernet II network frame.

The tagged frames contain a VLAN tag for assigning a virtual local area network (VLAN), dividing the structure into physical and logical layers. This means that using VLAN, subnets can be implemented without installing hardware. To identify frames in a VLAN, the Tag field is required. At the physical level, VLANs work through switches.

In the OSI model, VLAN operates at link layer 2 and controls the flow of data. With VLANs, networks can become more efficient by being subnetted. For information that the switch processes, marked frames are responsible. In the Tag field, it is inserted before the Type field and uses 4 bytes, which increases the minimum Ethernet frame size by 4 bytes.

802.3 Communication Sync

The synchronization process is performed by a bit stream, which allows the transmitter and receiver to synchronize the communication. The preamble is an alternating pattern of binary 56 units and zeros. The preamble is immediately followed by the separator of the initial frames:

1. The start separator is always 10101011 and is used to indicate the start of information.
2. The destination MAC address receiving the data. When the card (NIC) listens on the wire, checks this field for its own MAC.
3. The source MAC of the transmitting machine.
4. The length of the frame in bytes. Although this field can contain any value from 0 to 65,534, it rarely exceeds 1500 for most serial connections. Networks typically use serial devices to access the Internet.
5. Filling data (aka Payload).
6. Data is inserted here. This is the place where the IP header and other information are placed if IP is used over Ethernet. This field contains IPX information if IPX / SPX (Novell) is used.
7. The Ethernet 802 3 frame format section contains four specific fields: DSAP - access point to the destination service, SSAP - access source to the Poiont service, CTRL - control bits for Ethernet communication, NLI - network layer interface.
8. FCS - contains the frame check sequence, which is calculated using the cyclic redundancy check (CRC) and allows you to detect errors and reject the frame if it seems damaged.

This version of the 802.3 package, dubbed the "Ethernet 802.3", was released by Novell before the widespread adoption of IEEE 802.3 and the popular IPX / SPX, which unfortunately led to frequent confusion with the standard. Unlike the classic Ethernet II model, the frame determines the exact end of the bit sequence for SFD.

This identifies the data packet as the 802.3 standard for the recipient. 802.3raw frames do not contain a protocol identifier, as they can only be used for Novell IPX. In addition, the transmitted information always has a prefix of 2 bytes, which consist of units. This is the only way to distinguish a “raw” frame from others in the 802.3 family.

IEEE 802.3raw can only be used for IPX because there is no Ethernet frame field identifier. The name IEEE 802.3raw is also a bit misleading, as Novell used the name without involving IEEE when developing the frame.

Using this frame means extra work for the user, because compatibility issues between devices can occur. Since 1993, Novell has recommended the Ethernet 802.2 standard, which used IEEE 802.3 to avoid the possibility of confusion with the "raw" frame.

New Features - DSAP and SSAP

To standardize the structure of 802.3 introduced new features - DSAP and SSAP, which replace the type field and the control field containing the LLC frame. This version defines up to 256 compatible protocols with important information integrated in the data field. The new control field establishes a “logical connection” (LLC). This moment provides transparency of the procedures of media sharing and can control the information flow.

The DSAP, or access point to the destination service, is a 1-byte field that simply acts as a pointer to the memory buffer of the receiving station. It tells the NIC where to put the information. This functionality is crucial in situations where users use multiple protocol stacks. SSAP, or the source service access point, is similar to DSAP and indicates the source of the sending process.

The classic Ethernet 802.3 SNAP frame structure with a special feature is an SNAP field for defining more than 256 protocols. Their numbers provide 2 bytes, in addition, the manufacturer can integrate a unique identifier (3 bytes). Unlike its predecessors, SNAP also provides backward compatibility with Ethernet II.

DSAP, SSAP and Control are clearly defined. Thanks to the recently added space for 802.3 information, SNAP is extremely versatile and enables inter-protocol compatibility. Nevertheless, he has a little less space for actual data.

Classic Ethernet 802.3 frame structure with tags. The tag field contains important information for VLAN integration.

VLAN tags can also be set in the most popular IEEE 802.3 frame format. In this frame, the Tag field uses 4 bytes and is implemented before the length specification. The minimum frame size is now increased from 4 to 68.

Overview: Ethernet Frame Blocks. Pay attention to the table.

Data Link Header

The transmission header standard is “Offset 0-5: Recipient Address”. The first six bytes of the Ethernet frame size make up the destination address. It indicates to which adapter the data frame is sent. The destination address indicates a broadcast message that is read by the receiving devices. The first three destination bytes are assigned and depend on the IEEE provider. The destination format is the same in all Ethernet implementations.

The next six bytes make up the source address, the standard is “Offset 6-11: source address”. It indicates which adapter the message was sent from. , . Ethernet.

13 14 , , 32- CRC, DLC . 64 . , IEEE 802.2. LLC « » . , , LLC , .

SAP « 17: », , LLC. 802.2 43 1497 , , TCP/IP IPX, .

FCS: The last 4 bytes the adapter reads are a frame sequence or CRC. When the voltage on the wire returns to zero, the adapter checks the last 4 bytes that it received for the checksum generated through a complex polynomial. If the calculated checksum does not match the value in the frame, it is discarded and does not reach the memory buffers at the station.

Comparison of 802.11 vs 802.3

One of the differences between 802.3 Ethernet and 802.11 wireless frames is the frame size. 802.3 frames have a maximum Ethernet frame size of 1,518 bytes with a payload of 1,500 bytes. If 802.3 frames are labeled 802.1Q for VLAN and user priority, the maximum size of 802.3 is 1522 bytes with a data payload of 1504 bytes.

802.3 have only the source address (SA) and destination (DA) in the level 2 header. 802.11 have up to four fields in the MAC header and use only three fields of the MAC address (4 in the WDS environment).

The 802.11 standard is capable of transporting a fast Ethernet frame format with an MSDU payload. It is equal to 2304 bytes of the upper level. The MAC addressing used by 802.11 frames is much more complex than with Ethernet frames. Depending on whether the 802.11 traffic is upstream or downstream, the definition of each of the four MAC fields in the layer 2 header will change.

The Ethernet frame structure of IEEE 802.3 is today the most popular and widely used in the local area network. However, some protocols require more space for specific information. Therefore, there are IEEE 802.11 frame options that provide additional data blocks for specific information, including the SNAP extension and VLAN tag.