Simulation of connected devices

• US 7,571,088 B2
• Filed: 03/31/2006
• Issued: 08/04/2009
• Est. Priority Date: 03/31/2006
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Papaefstathiou, Efstathios, Peterson, Glenn R., Oslake, John M., Dournov, Pavel A.
• Primary Examiner(s): PHAN, THAI Q
• Application Number:
  US11/394,483
• Publication Number:
  US 20070233450A1
• Time in Patent Office: 1,222 Days
• Field of Search: 703/14, 703/21, 703/22, 703/16, 717/124
• US Class Current: 703/16
• CPC Class Codes: G06F 2111/08 Probabilistic or stochastic CAD
• CPC Class Codes: G06F 30/20 Design optimisation, verification or simulation optimisation, verification or simulation of circuit designs G06F30/30

Abstract

Simulating device interactions. A method may be practiced in a computing system for simulating interconnected devices. The method of simulating device interactions may be done in performing an overall transaction to obtain an output of system performance characteristics including measurement of latencies and/or device loads for actions performed by devices in performing the overall transaction. The method includes dividing a transaction into individual actions. The actions are applied to appropriate device models to produce latencies and/or device utilizations for the action as applied to the appropriate device model. This may be done by including an indication of an action type and optionally an action subservice. Memory resources may be conserved by ending the simulation when latencies and/or device utilizations have settled. Device utilization and latency may be aggregated and averaged over time. When changes in the latencies and/or device utilizations are within a predetermined range, the simulation can be ended.

Claims

1. In a computing system for simulating interconnected devices, a method of simulating device interactions in performing an overall transaction to obtain an output of system performance characteristics including measurement of at least one of latencies or device utilizations for actions performed by devices in performing the overall transaction, the method comprising:

dividing a transaction into individual actions comprising subparts of the transaction;

applying the individual actions to appropriate device models, by including an indication of an action type for each individual action, to produce at least one of latencies or device utilizations for the action as applied to the appropriate device model;

storing at least one averaged transaction latency or device utilization over time for one or more transactions; and

aggregating one or more current transaction latencies or device utilizations into the one or more stored averaged transaction latencies or device utilizations.

2. The method of claim 1 further comprising, obtaining at least one of a transaction latency or device utilizations by combining the latencies or device utilizations for the actions.

3. The method of claim 1, wherein applying the individual actions to the appropriate device models further includes an indication of an action subservice.

4. The method of claim 3, wherein one or more device models are hierarchical comprising sub-device models within the device model, the device model further comprising an indication of sub-device models assigned to actions based on the action type and subservice, the method further comprising at the device model assigning an action to a sub-device model by referencing the action type and subservice for an action.

5. The method of claim 4, wherein a plurality of sub-device models are assigned to a particular action based on the action type and subservice, wherein assigning an action to a sub-device model by referencing the action type and subservice for an action comprises assigning a sub-device model based on at least one of a round-robin assignment, a random assignment, a prioritized random assignment, or a least busy assignment.

6. The method of claim 1, wherein aggregating the at least one latency or device utilizations for the actions comprises adding latencies together for actions that are preformed serially.

7. The method of claim 1, wherein aggregating the at least one latency or device utilization for the actions comprises including a longest latency of two or more latencies when the two or more latencies are performed in parallel.

8. The method of claim 1, further comprising maintaining a fixed memory footprint when simulating device interactions by only maintaining at least one of a fixed number of stored averaged transaction latencies or device utilizations and at least one of a fixed number of current transaction latencies or device utilizations.

9. The method of claim 1, further comprising storing one or more averaged latencies or device utilizations for the actions.

10. The method of claim 1 further comprising:

if the stored averaged transaction latencies or device utilizations have statistical characteristics that are within a predetermined or otherwise computed range for such statistical characteristics, ending simulating the device interactions.

11. In a computing system for simulating interconnected devices, a computer readable storage medium including data structures for defining devices useful for simulating device interactions in performing an overall transaction to obtain an output of system performance characteristics including measurement of at least one latency or device utilizations for actions performed by devices in performing the overall transaction, the computer readable storage medium comprising:

a first field defining an action type for an action;

optionally a second field related to the first field defining a subservice defining a particular class of the action type;

one or more device model fields defining device models comprising methods for computing at least one the latencies or device utilizations for actions simulated by the device models; and

a third field comprising a scheduling key correlating a device model to an action type and optionally a subservice; and

computer-executable instructions for implementing a method comprising;

dividing a transaction into individual actions comprising subparts of the transaction;

applying the individual actions to appropriate device models, by including an indication of an action type for each individual action, to produce at least one of latencies or device utilizations for the action as applied to the appropriate device model;

storing at least one averaged transaction latency or device utilization over time for one or more transactions; and

aggregating one or more current transaction latencies or device utilizations into the one or more stored averaged transaction latencies or device utilizations.

12. The computer readable storage medium of claim 11, wherein the first and second field are defined within a master device model field defining a master device, the master device model field further comprises a scheduling policy for selecting other device model fields to apply an action to when multiple fields specify scheduling keys correlating different device models to the same action type and optionally subservice.

13. The computer readable storage medium of claim 12, wherein the scheduling policy is one or more of random, priority random, least busy or random, or round robin.

14. The computer readable storage medium of claim 11, wherein the scheduling key comprises a subservice key including action type and subservice parameters.

15. The computer readable storage medium of claim 11, wherein the scheduling key comprises a communication key including action type, communication peer, and communication direction.

16. The computer readable storage medium of claim 11, wherein the one or more device model fields defining device models that define methods for computing latencies for one or more of disk I/O operations, processor activities, or network transmissions.

17. The computer readable storage medium of claim 11, wherein one or more of the device model fields defining device models includes lower level device model components.

18. The computer readable storage medium of claim 11, wherein the computer-executable instructions for implementing a method further comprising:

if the stored averaged transaction latencies or device utilizations have statistical characteristics that are within a predetermined or otherwise computed range for such statistical characteristics, ending simulating the device interactions.