Clock synchronization using correlation events

• US 8,630,316 B2
• Filed: 06/19/2008
• Issued: 01/14/2014
• Est. Priority Date: 06/19/2008
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Haba, Erez
• Primary Examiner(s): Mered, Habte
• Assistant Examiner(s): MENSAH, PRINCE AKWASI
• Application Number:
  US12/142,748
• Publication Number:
  US 20090319819A1
• Time in Patent Office: 2,035 Days
• Field of Search: 370/250, 370/254, 370/401, 370/503, 370/510, 370/513, 370/514, 709/201, 709212-216, 709/230, 709/246, 709/25, 709300-302
• US Class Current: 370/510
• CPC Class Codes: G06F 1/12 Synchronisation of different clock signals provided by a plurality of clock generators
• CPC Class Codes: H04J 3/0661 using timestamps
• CPC Class Codes: H04L 41/142 using statistical or mathematical methods
• CPC Class Codes: H04L 43/106 using time related information in packets, e.g. by adding timestamps
• CPC Class Codes: H04N 21/242 Synchronization processes, e.g. processing of PCR [Program Clock References] arrangements for synchronising broadcast or distribution via plural systems in broadcast distribution systems H04H20/18

Abstract

Two clocks may be synchronized by calculating skew and offset values that may be determined from several correlation events. A correlation event may be the passing of messages in both directions between the two devices. The skew and offset values may be used to determine the time of non-correlated events. The clock synchronization may be performed on a real time basis or may be performed on a post processing basis. One method for calculating the skew and offset may use inequalities within a solution space to refine a solution set with multiple sets of correlation events.

Claims

1. At a device, a method implemented on a processor, said method comprising:

identifying a plurality of correlation events having an initiating timestamp and a receiving timestamp, said correlation events occurring between a first device having a first clock and a second device having a second clock, said plurality of correlation events comprising a first correlation event having said initiating timestamp correlated to said first clock and a second correlation event having said initiating timestamp correlated to said second clock;

determining that a solution set for an expression of the differences between said first clock and said second clock is not bounded, the solution set defining a linear relationship between said first clock and said second clock based on said first and second correlation events;

in response to determining that said solution set is not bounded;

identifying a further correlation event between said first device and said second device;

calculating a further linear inequality for said further correlation event;

and including said further linear inequality in said solution set;

determining that said solution set including said further linear inequality is a bounded solution set;

identifying a first event not being one of said plurality of correlation events, said first event having a third timestamp being defined by said second clock;

and using said bounded solution set to determine a fourth timestamp from said third timestamp, said fourth timestamp being with respect to said first clock.

2. The method of claim 1, wherein calculating a linear inequality for said correlation event comprises using a set of linear correlation constants to calculate said linear inequality, said of linear correlation constants comprising a skew constant and an offset constant.

3. The method of claim 1, at least one of said initiating timestamp and said receiving timestamp being determined by a transport layer service.

4. The method of claim 1, at least one of said initiating timestamp and said receiving timestamp being determined by a message passing interface.

5. The method of claim 1, said calculating being done by post processing.

6. The method of claim 5, said post processing comprising comparing a first log file from said first device and a second log file from said second device to identify at least one of said plurality of correlation events.

7. The method of claim 2, further comprising:

determining a single point within said bounded solution set to represent said set of linear correlation constants, said single point determined based on said linear inequalities included in said solution set.

8. The method of claim 7, said single point being determined based on a centroid of said bounded area.

9. The method of claim 7 further comprising:

determining an accuracy factor based on an area calculation of said bounded area.

15. The method of claim 1, further comprising:

identifying an additional correlation event between said first device and said second device;

calculating an additional linear inequality for said additional correlation event;

including said additional linear inequality in said solution set to refine said solution set;

and using said refined solution set to improve the accuracy of determining said fourth timestamp from said third timestamp.

18. The method of claim 1, further comprising:

calculating a linear inequality based on said first and second correlation events, said linear inequality defining said linear relationship between said first clock and said second clock;

and including said linear inequality in said solution set.

10. A system comprising:

a first clock;

a first communication mechanism configured to send and receive messages and include a timestamp on each of said messages using said first clock;

a clock synchronization mechanism configured to synchronize a second clock from a second device with said first clock by a method comprising;

identifying a plurality of said messages transmitted to said second device and received from said second device, said plurality of messages having a timestamp from said second clock;

determining that a solution set for an expression of the differences between said first clock and said second clock is not bounded, the solution set defining one or more linear relationships between said first clock and said second clock based on said identified plurality of messages;

in response to determining that said solution set is not bounded;

identifying a further message involving said second device;

calculating a further linear inequality for said further message;

and including said further linear inequality in said solution set;

determining that the solution set including said further linear inequality is a bounded solution set;

and using said bounded solution set to estimate synchronization between said first clock and said second clock.

11. The system of claim 10, said clock synchronization mechanism further configured to calculate said set of linear correlation constants in real time.

12. The system of claim 11, said clock synchronization mechanism further configured to:

determine a single point within said bounded area to represent said set of linear correlation constants.

16. The system of claim 10, wherein said method further comprises identifying an additional message sent between said first device and said second device;

calculating an additional linear inequality for said additional message;

including said additional linear inequality in said solution set to refine said solution set;

and using said refined solution set to more accurately estimate synchronization between said first clock and said second clock.

19. The system of claim 10, further comprising:

calculating one or more linear inequalities based on said plurality of messages, said one or more linear inequalities defining said one or more linear relationships between said first clock and said second clock; and

including said one or more linear inequalities in said solution set.

13. A computer storage device comprising computer executable instructions configured to perform a method comprising:

reading a log file to identify a plurality of correlation events having an initiating timestamp and a receiving timestamp, said correlation events occurring between a first device having a first clock and a second device having a second clock, said plurality of correlation events comprising a first correlation event having said initiating timestamp created by said first device and a second correlation event having said receiving timestamp created by said second device;

determining that a solution set for an expression of the differences between said first clock and said second clock is not bounded, the solution set defining a linear relationship between said first clock and said second clock based on said first and second correlation events;

in response to determining that said solution set is not bounded;

reading said log file to identify a further correlation event between said first device and said second device;

calculating a further linear inequality based on said further correlation event;

and including said further linear inequality in said solution set;

determining that said solution set including said further linear inequality is a bounded solution set;

and using said bounded solution set to estimate synchronization between said first clock and said second clock.

14. The computer storage device of claim 13, wherein using said bounded solution set to estimate synchronization between said first clock and said second clock comprises determining a centroid for said bounded solution set.

17. The computer storage device of claim 13, wherein said method further comprises:

identifying an additional correlation event between said first device and said second device;

calculating an additional linear inequality for said additional correlation event;

including said additional linear inequality in said solution set to refine said solution set by reducing the area of said solution set;

and using said refined solution set to more accurately estimate synchronization between said first clock and said second clock.

20. The computer storage device of claim 13, further comprising:

calculating said linear inequality based on said first and second correlation events, said linear inequality defining said linear relationship between said first clock and said second clock;

and including said linear inequality in said solution set.