Packet Framing

Packet framing defines the start and end of a data packet, allowing the receiver to synchronize and correctly extract the payload from a continuous bit stream. Common methods include:

• Character-based framing (e.g., sentinel characters like STX/ETX).
• Bit-oriented framing (e.g., HDLC's use of the '01111110' flag sequence).
• Length-based framing, where a header field specifies the packet's total size.

A frame's header contains control information for delivery, such as source/destination addresses, sequence numbers, and protocol identifiers. The trailer, often containing a Frame Check Sequence (FCS) like a CRC, is used for error detection. This structure is visible in protocols like Ethernet (IEEE 802.3), where the frame comprises a preamble, MAC addresses, EtherType, payload, and FCS.

The core function of framing is to encapsulate the Protocol Data Unit (PDU) from a higher layer (e.g., an IP packet) within the frame's data field. This creates a nesting model where each layer adds its own header/trailer, a principle central to the OSI and TCP/IP models. The maximum payload size is defined by the Maximum Transmission Unit (MTU).

To prevent the accidental appearance of control sequences (like frame delimiters) within the payload, bit stuffing is used. In HDLC, a '0' bit is inserted after every five consecutive '1's in the data stream. The receiver destuffs these bits to recover the original data, ensuring data transparency where any bit pattern can be transmitted.

Framing enables link-layer reliability mechanisms. Error control uses the trailer's FCS to detect corrupted frames, which are then discarded. Flow control manages the data rate between sender and receiver to prevent overflow, using frame headers for acknowledgments and window sizing, as seen in protocols like HDLC and PPP.

Different link-layer technologies implement distinct framing standards:

• Ethernet (IEEE 802.3): Uses preamble and SFD for synchronization, with MAC addresses in the header.
• Point-to-Point Protocol (PPP): Employs a flag byte (0x7E) for delimitation and a protocol field to identify the encapsulated packet.
• High-Level Data Link Control (HDLC): The canonical bit-oriented framing protocol using flag sequences and bit stuffing.

Uses special control characters (like ASCII's STX/ETX) to mark packet boundaries. The Data Link Escape (DLE) character is often used to 'escape' control characters appearing in the data itself, a technique called character stuffing.

• Example: The Binary Synchronous Communication (BISYNC) protocol.
• Limitation: Not suitable for binary data where any byte value can appear.

Defines a unique flag sequence (e.g., 01111110) to delimit frames. To prevent this flag from appearing in the data, bit stuffing is used: a 0 is inserted after any sequence of five consecutive 1s in the data stream, which the receiver removes.

• Example: High-Level Data Link Control (HDLC) and its derivative, Point-to-Point Protocol (PPP).
• Advantage: Efficient and transparent for any data pattern.

Encodes the packet length in a fixed-size header field. The receiver reads the length field first, then knows exactly how many subsequent bytes belong to that frame. This requires a maximum transmission unit (MTU) to be defined.

• Example: Ethernet (IEEE 802.3) frames use a 2-byte length/type field.
• Challenge: A corrupted length field can cause catastrophic de-synchronization.

Relies on timing gaps or specific clock signals between packets instead of in-band delimiters. Often used in conjunction with other methods or in synchronous communication systems where the data stream is continuous.

• Example: SONET/SDH optical networks use a fixed frame duration (125 µs).
• Use Case: Ideal for real-time, isochronous data like voice and video streams.

Uses invalid signal encodings from the physical layer to mark frame boundaries. Since these signals cannot appear in valid data, they provide unambiguous delimiters without data alteration.

• Example: Manchester encoding violations in some legacy LANs.
• Advantage: Simplifies receiver logic as delimiting is handled at the lowest layer.

Framing is not explicitly signaled; boundaries are inferred from higher-layer protocol semantics or fixed packet sizes. Common in cell-based networks.

• Example: Asynchronous Transfer Mode (ATM) uses fixed 53-byte cells.
• Example: Some serial protocols assume every N bytes is a complete message.

Packet framing operates primarily at the Data Link Layer (Layer 2) of the OSI model, which is responsible for node-to-node data transfer and error detection. It sits above the Physical Layer and below the Network Layer, framing raw bit streams from the physical medium into structured packets for higher-level protocols.

The MTU is the largest size of a protocol data unit (like a packet or frame) that can be transmitted in a single network transaction. Exceeding the MTU requires fragmentation. In blockchain P2P networks, MTU limits influence how large blocks or transaction batches are packaged for transmission between nodes.

A PDU is a specific block of data transferred at a given OSI layer. The naming changes per layer:

• Segment at Transport Layer (Layer 4)
• Packet at Network Layer (Layer 3)
• Frame at Data Link Layer (Layer 2)
• Bits at Physical Layer (Layer 1) Framing is the process of encapsulating a Layer 3 packet into a Layer 2 frame by adding header and trailer information.

These are techniques to ensure frame delimiters (like the flag sequence 01111110) do not appear in the actual data payload, which would cause premature frame termination.

• Bit Stuffing: Inserts a 0 after five consecutive 1s in the data stream.
• Byte Stuffing (Character Stuffing): Uses special escape characters (like DLE - Data Link Escape) before a reserved byte in the data.

Framing methods differ based on timing synchronization:

• Asynchronous Framing: Used in serial communications like UART. Each character is framed individually with start and stop bits. Inefficient for high-speed data.
• Synchronous Framing: Transmits data in continuous, synchronized blocks. Used in protocols like HDLC and PPP. More efficient and is the standard for modern network and blockchain P2P communication.

Frame trailers typically contain error-checking codes to ensure data integrity during transmission. Common methods include:

• Cyclic Redundancy Check (CRC): A powerful algorithm that computes a checksum based on the frame's content.
• Frame Check Sequence (FCS): The field in the trailer that holds the CRC result. The receiving node recalculates the CRC; a mismatch indicates a corrupted frame, triggering a retransmission request.