Granted Patent B2

US 7,839,781 · App. 12/210,087 · Granted Nov 23, 2010

Abstract

Claims (39)

1. A method for configuring service curves for managing the output port of a networking device having a link rate, the method comprising:

configuring, according to user input, one or more traffic classes, wherein each class is assigned a bandwidth guarantee; and

computing a service curve for each of the one or more traffic classes, each service curve being a function of the bandwidth guarantee assigned to that class,

wherein at least one service curve is shifted by an amount of time, and

wherein the sum of all bandwidth guarantees is less than or equal to the link rate.

2. The method of claim 1 , further comprising:

assigning a priority category to each traffic class, wherein the priority category is associated with a burst parameter that defines the amount of time, during which an active traffic class is allowed to pre-empt traffic classes having lower priority categories.

3. The method of claim 2 , wherein the burst parameter further defines a partition of a time-axis into adjacent non-overlapping time zones that represent a plurality of priority categories available to the one or more traffic classes.

4. The method of claim 2 , wherein the burst parameter is defined in units of time.

5. The method of claim 1 , wherein the service curve is a two-piece linear function having a first slope and a second slope.

6. The method of claim 1 , wherein the act of configuring one or more traffic classes comprises:

receiving a delay priority associated with an active traffic class, the delay priority defining the amount of time the active traffic class can pre-empt traffic classes having lower delay priorities.

7. The method of claim 1 , wherein the user input comprises information data about the bandwidth guarantee and a priority category associated with each traffic class.

8. The method of claim 1 , wherein the amount of time is an average time period for transmitting a number of data packets.

9. A method for mapping sharing policies onto a link sharing class hierarchy, the method comprising:

constructing the link sharing class hierarchy having a plurality of sharing classes;

assigning a sharing priority level to each of the plurality of sharing classes;

allocating a portion of a link bandwidth to a first sharing priority level;

servicing at least one traffic class associated with the first sharing priority level, which is the highest priority level; and

dynamically allocating excess service to sharing classes having a second sharing priority level, which is lower than the first sharing priority level;

wherein each of the plurality of sharing classes includes a service curve having a linear function;

wherein a root of the class hierarchy includes a directed edge to at least one of the plurality of sharing classes.

10. The method of claim 9 , wherein the act of constructing the link sharing class hierarchy comprises:

computing a directed acyclic graph having the link bandwidth as the root.

11. The method of claim 9 , wherein the sharing classes having a second sharing priority level are “overflow” classes that, in the event that the excess service is not used, distribute the excess service to other active classes.

12. The method of claim 11 , wherein the distribution of the excess service is in proportion to a bandwidth guarantee associated with each of the active classes.

13. The method of claim 9 , wherein the dynamically allocating excess service to traffic classes having a second sharing priority level is in proportion to their respective assigned bandwidth guarantees without being first allocated to an overflow class.

14. A service curve scheduler that configures service curves for managing an output port having a link rate, the scheduler comprising:

a module adapted to configure one or more traffic classes based on user input, each traffic class being associated with a bandwidth guarantee; and

a module adapted to compute a service curve for each of the one or more traffic classes, each service curve being a function of the bandwidth guarantee, wherein at least one service curve is shifted by an amount of time, and wherein the sum of all bandwidth guarantees is less than or equal to the link rate.

15. The service curve scheduler of claim 14 , further comprising:

a module adapted to assign a priority category to each traffic class, wherein the priority category is associated with a burst parameter that defines the amount of time, during which an active traffic class is allowed to pre-empt traffic classes having lower priority categories.

16. The service curve scheduler of claim 14 , wherein the burst parameter further defines a partition of a time-axis into adjacent non-overlapping time zones that represent a plurality of priority categories available to the one or more traffic classes.

17. The service curve scheduler of claim 14 , wherein the burst parameter is defined in units of time.

18. The service curve scheduler of claim 14 , wherein the service curve is a two-piece linear function having a first slope and a second slope.

19. The service curve scheduler of claim 14 , further comprising;

a module adapted to receive a delay priority associated with an active traffic class, the delay priority defining the amount of time the active traffic class can pre-empt traffic classes having lower delay priorities.

20. The service curve scheduler of claim 14 , wherein the user input comprises information data about the bandwidth guarantee and a priority category associated with each traffic class.

21. The service curve scheduler of claim 14 , wherein the amount of time is an average time period for transmitting a number of data packets.

Assignments (19)