Processing network packets using fractional time stamp values

• US 7,751,322 B1
• Filed: 03/12/2008
• Issued: 07/06/2010
• Est. Priority Date: 06/23/2004
• Status: Expired due to Fees

• Current Assignee: Dinochip Incorporated
• Original Assignee: Dinochip Incorporated
• Inventors: Chen, Li-Sheng, Shao, Huadong, Nguyen, Dziem Dinh, Tang, Qian-Yu
• Primary Examiner(s): Ngo; Ricky
• Assistant Examiner(s): PATEL, CHANDRAHAS B
• Application Number:
  US12/046,638
• Time in Patent Office: 846 Days
• Field of Search: 370/230, 709/224
• US Class Current: 370/230
• CPC Class Codes: H04L 43/087 Jitter
• CPC Class Codes: H04L 43/106 using time related information in packets, e.g. by adding timestamps

Abstract

Fractional or fuzzy time stamp values are used during processing network traffic. Handling fractional time stamp values permits scaling down of time stamp values, so that fewer bits, and memory, are needed to store each fractional time stamp value. According to a specific technique, departure times are set based on a quantity X plus Y/Z, where X is an integer, Y is a numerator, and Z is a denominator. The departure times are synchronized with the ideal departure times after (Z*X+Y) time units. A technique is provided to reduce cell delay variation.

Claims

1. A method for data networking comprising:

providing a first connection and a second connection, each capable of connecting to pass data to a network port;

providing a first fractional time stamp expression for the first connection; and

allocating the network port to the first connection for X blocks of time based on the first fractional time stamp expression, wherein X is at least two, a first of the X blocks has M time units, a second of the X blocks has N time units, and M is not equal to N, and

a number of bits used to represent the fractional time stamp expression will be a logarithm, base 2, of a bandwidth of the network port divided by a selected granularity.

2. The method of claim 1 comprising:

providing a second fractional time stamp expression for the second connection.

3. The method of claim 1 wherein one time unit comprises one cell size divided by a bandwidth of the network port.

4. The method of claim 1 wherein the port is at least one of a DS-3, OC-3, OC-12, or OC-48 port connection.

5. The method of claim 1 further comprising:

sorting the first and second connection using the fractional time stamp value.

6. The system of claim 1 wherein a first time the first connection is connected to the port for a time period different from that for a second time the first connection is connected to the port.

7. The method of claim 1 wherein the fractional time stamp expression is derived by scaling down a corresponding nonfractional time stamp value by a user-selected factor.

8. The method of claim 7 wherein the fractional time stamp value is represented using fewer binary bits than the corresponding nonfractional time stamp value.

9. A system comprising:

a port having a first bandwidth;

a plurality of connections, each capable of connecting to pass data to the port; and

a controller, wherein the controller connects a first connection, having a second bandwidth, of the connections to the port for a time period which is based on a fractional time stamp value calculated using time value plus the first bandwidth divided by the second bandwidth, and a number of bits used to represent the fractional time stamp expression will be a logarithm, base 2, of a bandwidth of the port divided by a selected granularity.

10. The system of claim 9 wherein the port is at least one of a DS-3, OC-3, OC-12, or OC-48 port connection.

11. The system of claim 9 wherein a first time the first connection is connected to the port for a time period different from that for a second time the first connection is connected to the port.

12. The system of claim 9 where the controller performs sorting using the fractional time stamp value.

13. The system of claim 9 wherein the fractional time stamp value is derived by scaling down a corresponding nonfractional time stamp value by a user-selected factor.

14. The system of claim 13 wherein the fractional time stamp value is represented using fewer binary bits than the corresponding nonfractional time stamp value.

15. A method for data networking comprising:

providing a first connection and a second connection, each capable of connecting to pass data to a network port;

providing a scale-down factor Z;

providing a first time stamp expression for the first connection;

scaling down the first time stamp expression using the scale-down factor Z, wherein the first time stamp expression becomes fractional and includes a quantity X plus Y/Z, where X is an integer, Y is a numerator, and Z is a denominator; and

allocating the network port to the first connection for Z time periods, wherein at least Y of the Z time periods has X time units and one of the Z time periods has (X+Y) time units.

16. The method of claim 15 wherein a value for Z is provided by a user.

17. The method of claim 15 further comprising:

sorting the first and second connection using the fractional time stamp value.

18. The method of claim 15 comprising:

scaling down the second time stamp expression using the scale-down factor Z.

19. The method of claim 15 wherein the Z time periods has a total of (Z*X+Y) time units.

20. The method of claim 15 wherein the second time stamp expression is not fractional.