Granted Patent B2

US 6,851,084 · App. 10/310,514 · Granted Feb 1, 2005

Abstract

Claims (21)

1. A method of correcting errors in a packet based communications system, wherein a source station transmits data packets to a destination station via network media, comprising:

compiling a serial data stream into a sequence of data packets for transmission to the destination station via the network media;

generating a first parity packet as a bit-wise exclusive-or of a first data packet and a second data packet from the sequence of data packets;

generating a second parity packet as a bit-wise exclusive-or of a third data packet and a fourth data packet from the sequence of data packets;

generating a third parity packet as a bit-wise exclusive-or of the first data packet and the third data packet from the sequence of data packets; and,

generating a fourth parity packet as a bit-wise exclusive-or of the second data packet and the fourth data packet from the sequence of data packets;

forming a code word from the first, second, third, and fourth parity packets, and the first, second, third, and fourth data packets;

associating a sequence number with each packet representative of a relative position of the packet in the code word;

transmitting the eight packet code word to the destination station via the network media;

receiving the eight packet code word at the destination station, and examining the sequence number associated with each packet to determine if one or more packets from the eight packet code word is missing, and identifying the one or more missing packets if any are missing;

processing the data packets and the parity packets within the associated code word at the destination station, according to the combinations of data packets used to generate the parity packets, so as to recover one or more missing packets, wherein processing the data packets includes:

first, correcting a single error, if it exists, in a parity group consisting of the first parity packet, the first data packet and the second data packet, and correcting a single error, if it exists, in a parity group consisting of the second parity packet, the third data packet and the fourth data packet; and

second, correcting a single error, if it exists, in a parity group consisting of the third parity packet, the first data packet and the third data packet, and correcting a single error, if it exists, in a parity group consisting of the fourth parity packet, the second data packet and the fourth data packet.

2. A method according to claim 1 , wherein processing the data packets further includes determining if errors exist after completing the second step, and if so, repeating the first step and the second step.

3. A system for correcting errors in a packet based communications system, wherein a source station transmits data packets to a destination station via network media, comprising:

a packet assembler in the source station for compiling a serial data stream into a sequence of data packets for transmission to the destination station via the network media;

an encoder in the source station (i) for receiving the data packets from the packet assembler, (ii) for generating a first parity packet as a bit-wise exclusive-or of a first data packet and a second data packet from the sequence of data packets; (iii) for generating a second parity packet as a bit-wise exclusive-or of a third data packet and a fourth data packet from the sequence of data packets; (iii) for generating a third parity packet as a bit-wise exclusive-or of the first data packet and the third data packet from the sequence of data packets; (iv) for generating a fourth parity packet as a bit-wise exclusive-or of the second data packet and the fourth data packet from the sequence of data packets. (v) for forming a code word from the first, second, third, and fourth parity packets, and the first, second, third, and fourth data packets, and (vi) for associating a sequence number with each packet representative of a relative position of the packet in the code word;

a decoder for receiving the eight packet code word at the destination station, and examining the sequence number associated with each packet to determine if one or more packets from the eight packet code word is missing, and identifying the one or more missing packets if any are missing, and for processing the data packets and the parity packets within the associated code word at the destination station, according to the combinations of data packets used to generate the parity packets, so as to recover one or more missing packets, wherein the decoder processes the data packets and the parity packets at the destination station by:

first, correcting a single error, if it exists, in a parity group consisting of the first parity packet, the first data packet and the second data packet, and correcting a single error, if it exists, in a parity group consisting of the second parity packet, the third data packet and the fourth data packet; and

second, correcting a single error, if it exists, in a parity group consisting of the third parity packet, the first data packet and the third data packet, and correcting a single error, if it exists, in a parity group consisting of the fourth parity packet, the second data packet and the fourth data packet.

4. A system according to claim 3 , wherein the decoder determines if errors exist after completing the second step, and if so, repeating the first step and the second step.

Assignments (14)