Method of transmitting data

• US 20030086413A1
• Filed: 08/30/2002
• Published: 05/08/2003
• Est. Priority Date: 08/31/2001
• Status: Active Grant

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Banchs, Albert, Tartarelli, Sandra
• Granted Patent: US 7,599,296 B2
• Time in Patent Office: Days
• US Class Current: 370/352
• CPC Class Codes: H04L 43/00 Arrangements for monitoring or testing data switching networks
• CPC Class Codes: H04L 43/0829 Packet loss
• CPC Class Codes: H04L 43/0864 Round trip delays
• CPC Class Codes: H04L 43/0888 Throughput
• CPC Class Codes: H04L 43/16 Threshold monitoring
• CPC Class Codes: H04L 47/10 Flow control; Congestion control
• CPC Class Codes: H04L 47/32 by discarding or delaying data units, e.g. packets or frames

Abstract

A method of transmitting data from customers (C1, C2, C3, C4, C5, C6, C7, C8, C9, C10) over a computer network, in particular over the Internet, where the data to be sent is split into packets, in particular into IP packets, where each packet is marked by one of at least two states (IN, OUT) and where the states (IN, OUT) determine which packets are dropped first, if packets are dropped during transmission, is, with regard to optimizing the drop rate of the packets, characterized in that the marking of the packet with a state of high drop precedence (OUT) is based on a random probability (p).

Claims

1. Method of transmitting data from customers (C1, C2, C3, C4, C5, C6, C7, C8, C9, C10) over a computer network, in particular over the Internet, where the data to be sent is split into packets, in particular into IP packets, where each packet is marked by one of at least two states (IN, OUT) and where the states determine which packets are dropped first, if packets are dropped during transmission, characterized in that the marking of the packet with a state (OUT) of high drop precedence is based on a random probability (p).

2. Method according to claim 1, characterized in that the marking of the packet with a high drop precedence (OUT) is based on a single random probability (p).

3. Method according to claims 1 and 2, characterized in that the probability (p) is measured for the traffic of each customer (C1, C2, C3, C4, C5, C6, C7, C8, C9, C10).

4. Method according to claims 1 to 3, characterized in that the computers of the network are linked with each other.

5. Method according to claim 4, characterized in that several customers (C1, C2, C3, C4, C5, C6, C7, C8, C9, C10) share at least parts of a link, in particular a line and/or a wireless connection or similar.

6. Method according to claims 4 or 5, characterized in that the links have maximum bandwidths and/or the customers have been provided an assigned maximum bandwidth (CIR) for the purpose of data transmission.

7. Method according to claim 6, characterized in that the traffic of the customer (C1, C2, C3, C4, C5, C6, C7, C8, C9, C10) is measured.

8. Method according to claims 6 or 7, characterized in that a packet is dropped when the assigned maximum bandwidth (CIR) of the customer (C1, C2, C3, C4, C5, C6, C7, C8, C9, C10) is exceeded during the transmission and/or the maximum bandwidth of the connection in not sufficient to transmit the packets.

9. Method according to claims 6 to 8, characterized in that the marking of the packet is based on the comparison of the current bandwidth with the assigned maximum bandwidth (CIR).

10. Method according to claim 9, characterized in that the comparison of the current bandwidth with the assigned maximum bandwidth (CIR) is based on a token bucket.

11. Method according to one of the claims 6 to 10, characterized in that the packet is marked with a high drop precedence if the current bandwidth is higher than the assigned maximum bandwidth (CIR).

12. Method according to one of the claims 1 to 11, characterized in that the sending of the data is done via the TCP transport protocol and in particular via TCP/IP.

13. Method according to one of the claims 1 to 12, characterized in that the packets are forwarded in a DiffServ environment, preferably via PHB, in particular with Assured Forwarding (AF) with WRED.

14. Method according to one of the claims 1 to 13, characterized in that the marking is done by means of two states.

15. Method according to claims 1 to 14, characterized in that the probability (p) at a given step is expressed as

p=k1×

(b_ref−

b)−

k2×

(b_ref−

b_old)+p_old

and that at the next step p_old is set equal to p and b_old is set equal to b.

16. Method according to one of the claims 1 to 15, characterized in that the probability (p) is compared to a random number (u) evenly distributed between 0 and 1.

17. Method according to claim 16, characterized in that the packet is marked with a high drop precedence (OUT) if the probability (p) is greater than the random number (u).

18. Method according to one of the claims 1 to 17, characterized in that the packets are put into a buffer preferably assigned to the core node.

19. Method according to one of the claims 1 to 18, characterized in that, in case of packet congestion during transmission, the packets marked with a high drop precedence (OUT) are discarded.

20. Method according to one of the claims 1 to 19, characterized in that the change in the TCP window size (W) is expressed as

21. Method according to one of the claims 1 to 20, characterized in that the change in the token bucket occupancy (b) is expressed as

22. Method according to claim 20 and/or claim 21, characterized in that the change in the TCP window size (W) and/or the token bucket occupancy (b) is linearized in the operation point, preferably at constant round trip time (RTT) and/or constant number of TCP sources (N)

23. Method according to one of the claims 1 to 22, characterized in that, assuming

24. Method according to one of the claims 1 to 23, characterized in that the token bucket occupancy (b) is stabilized by way of a controller, especially a PI controller

25. Method according to claim 24, characterized in that the control system constant is set to a value greater than the TCP time constant and especially that the zero of the controller is set to

26. Method according to claims 24 or 25, characterized in that especially by invoking the Nyquist criterion the gain (K) in controller is set to

27. Method according to claims 15, 25 and 26, characterized in that k1 is computed by a preferably bilinear transformation as

28. Method according to claims 15, 25 and claim 26 or claim 27, characterized in that k2 is computed by a preferably bilinear transformation as