Site home page
(news and notices)

Get alerts when Linktionary is updated

Book updates and addendums

Get info about the Encyclopedia of Networking and Telecommunicatons, 3rd edition (2001)

Download the electronic version of the Encyclopedia of Networking, 2nd edition (1996). It's free!

Contribute to this site

Electronic licensing info

 

 

Framing in Data Transmissions

Related Entries    Web Links    New/Updated Information

  
Search Linktionary (powered by FreeFind)

Note: Many topics at this site are reduced versions of the text in "The Encyclopedia of Networking and Telecommunications." Search results will not be as extensive as a search of the book's CD-ROM.

A point-to-point connection between two computers or devices consists of a wire in which data is transmitted as a stream of bits. However, these bits must be framed into discernible blocks of information. Framing is a function of the data link layer. It provides a way for a sender to transmit a set of bits that are meaningful to the receiver. Ethernet, token ring, frame relay, and other data link layer technologies have their own frame structures. Frames have headers that contain information such as error-checking codes.

There are three different types of framing, each of which provides a way for the sender to tell the receiver where the block of data begins and ends:

  • Byte-oriented framing    Computer data is normally stored as alphanumeric characters that are encoded with a combination of 8 bits (1 byte). This type of framing differentiates one byte from another. It is an older style of framing that was used in the terminal/mainframe environment. Examples of byte-oriented framing include IBM's BISYNC protocol.

  • Bit-oriented framing    This type of framing allows the sender to transmit a long string of bits at one time. IBM's SDLC (Synchronous Data Link Control) and HDLC (High-level Data Link Control) are examples of bit-oriented protocols. Most LANs use bit-oriented framing. There is usually a maximum frame size. For example, Ethernet has a maximum frame size of 1,526 bytes. The beginning and end of a frame is signaled with a special bit sequence (01111110 for HDLC). If no data is being transmitted, this same sequence is continuously transmitted so the end systems remain synchronized.

  • Clock-based framing    In a clock-based system, a series of repetitive pulses are used to maintain a constant bit rate and keep the digital bits aligned in the data stream. SONET (Synchronous Optical Network) is a synchronous system in which all the clocks in the network are synchronized back to a master clock reference. SONET frames are then positioned within the clocked stream.

The advantage of using frames is that data is broken up into recoverable chunks that can easily be checked for corruption. A glitch in the line during the transmission will corrupt some frames. Only the lost frames and not the entire set of data needs to be retransmitted. Detecting and correcting errors is discussed under "Error Detection and Correction."

The general format for frames is a header followed by a data payload area. The frame size is usually fixed, while the data area is variable. Most data link networks use variable-size framing, which has advantages (more efficient use of the network) and disadvantages (unpredictable traffic flows and the inability to provide quality of service).

An Ethernet frame (IEEE 802.3), as pictured in Figure F-16, can hold up to 1,500 bytes of data. Note that the frame includes header and trailer bytes along with the encapsulated data. The information in the header adds overhead in the form of additional data that must be transmitted.

ANCHOR HERE: Figure 16 (see book)

The preamble contains a set of bits that help the sender and receiver synchronize their transmissions. The Ethernet bit pattern is 10101010, which produces a 10-MHz square wave clocking signal for 5.6 microseconds. The SFD (start frame delimiter) indicates the actual start of the frame. The destination and source addresses are self-explanatory. The LEN (length) field indicates the length of the data portion of the frame. PAD (padding) is used for fill-in bits. CRC (cyclic redundancy checksum) is used for error checking.

Framing is specific to the data link layer. Upper-layer protocols also divide data into discernible "packets" of information, but the terminology used to define packets at each layer is different, as outlined next. Note that these descriptions assume the Internet protocol suite.

  • Message    The actual application data, command, or instruction specified by the user or application. A message is encapsulated within a TCP segment assuming TCP is used.

  • Segment    The packet of information exchanged between two connected systems (peers) that contains transport layer protocol information. TCP exchanges segments. Segments encapsulate upper-layer messages.

  • Datagram    The packet of information that is exchanged between two connected systems that contains network layer protocol information. IP exchanges datagrams. Datagrams encapsulate segments.

  • Frame    The packet of information at the data link layer with a layout that is specific to the protocol in use. Frames encapsulate datagrams.

Figure F-17 illustrates how data is passed down the protocol stack and encapsulated by the protocols at each layer. The TCP/IP protocol is illustrated in this example. At the transport layer, TCP encapsulates upper-layer data and adds its own header to create a segment. At the network layer, IP encapsulates the segment and its own header into a datagram. Finally, the data link layer encapsulates the datagram into a frame or frames. Note that datagrams may be fragmented and placed into multiple frames. Refer to "Network Architecture" and "Fragmentation and Reassembly" for more information.

ANCHOR HERE: Figure 17 (see book)

Framing is handled by the underlying network protocols and equipment. Ethernet, token ring, frame relay, ATM, and other data link technologies use their own framing standards (ATM frames are actually called "cells" and are a fixed size). Upper-layer protocols are generally designed to interface with any of these protocols and framing methods. When building a network application, you generally don't need to care about which underlying LAN will be used, as long as you ensure the application is compatible with a network protocol suite like TCP/IP.

Note that frames are confined to specific network segments. An internetwork consists of multiple data link segments, as pictured in Figure F-18. One segment may be Ethernet while another is token ring. The connecting bridge or router must remove the encapsulated data from the frames it receives and reencapsulate the data into the frame type of the next network. In some cases, the next network might require a smaller frame size, so the device might need to fragment the data before encapsulating it into frames. Note that once data is fragmented, it travels that way all the way to the final destination.

ANCHOR HERE: Figure 18 (see book)

Alteon has proposed expanding the Ethernet frame size from its 1,500-byte size to a "jumbo frames" size of 9,000 bytes. The company claims that jumbo frames can provide a 300-percent increase in throughput for Gigabit Ethernet networks by reducing the number of frames that must be processed. Alteon believes that the 9,000-byte size (and not a larger size) is a good balance between efficiency and the processing requirements of error checking and handling large blocks of data.




Copyright (c) 2001 Tom Sheldon and Big Sur Multimedia.
All rights reserved under Pan American and International copyright conventions.