Modified machine architecture with advanced synchronization

• US 8,122,200 B2
• Filed: 01/23/2008
• Issued: 02/21/2012
• Est. Priority Date: 10/25/2005
• Status: Active Grant

• Current Assignee: Waratek Limited (Waratek Holdings Ltd.)
• Original Assignee: Waratek Pty Limited
• Inventors: Holt, John M.
• Primary Examiner(s): Gu, Shawn X
• Application Number:
  US12/011,218
• Publication Number:
  US 20080215701A1
• Time in Patent Office: 1,490 Days
• Field of Search: None
• US Class Current: 711/152
• CPC Class Codes: G06F 12/0815 Cache consistency protocols
• CPC Class Codes: G06F 9/526 Mutual exclusion algorithms

Abstract

A multiple computer environment is disclosed in which an application program executes simultaneously on a plurality of computers (M1, M2, . . . Mn) interconnected by a communications network (53) and in which the local memory of each computer is maintained substantially the same by updating in due course. A lock mechanism is provided to permit exclusive access to an asset, object, or structure (ie memory location) by acquisition and release of the lock. In particular, before a new lock can be acquired by any other computer on a memory location previously locked by one computer, any re-written content(s) for the previously locked memory location are transmitted to all the other computers and their corresponding memory locations (before the in due course updating). Thus when the new lock is acquired all the corresponding memory locations of all computers have been updated.

Claims

1. A method carried out in a multiple computer system comprising a plurality of machines each including a processor and a local memory coupled with the processor including a first machine and a second machine interconnected via a communications link or network operating in a replicated shared memory arrangement, the method comprising:

a. acquiring a lock on the first machine having a replicated memory location;

b. recording an identifier of the replicated memory location written-to with an updated content prior to releasing said lock on the first machine;

c. releasing said lock on the first machine; and

d. prior to permitting an operation of the lock by the second machine, or prior to said second machine acquiring the lock, transmitting the recorded identifier and the updated content to said second machine.

2. A method as in claim 1, wherein the transmitting of the recorded identifier and the updated content to said second machine is performed so that said second machine upon operating on said lock has already received said updated content.

3. A method as in claim 2, wherein the step (d) of transmitting of the recorded identifier and the updated content to said second machine is performed as part of in due course updating so that said second machine upon operating on said lock has already received said updated content comprises transmitting the recorded identifier and contents of the written-to the replicated memory location of step (b) to the second machine.

4. A method as in claim 3, wherein said in due course updating includes updating replicated memory locations so that they are updated to remain substantially similar upon occasion of any one of said plurality of machines simultaneously operating an application program causing modification of the contents of said replicated memory location.

5. A method as in claim 3, wherein the transmitting of step (d) comprises transmitting only the final content of each written-to memory location of step (b).

6. A method as in claim 3, wherein the transmitting of step (d) comprises transmitting all updated contents of each written-to memory location including a final updated content and an intermediary changed content.

7. A method as in claim 1, wherein the transmitting of the recorded identifier and the updated content to said second machine is performed so that said second machine upon operating on said lock has already received said updated content without waiting for ordinary updating of said written-to replicated memory location of all said plurality of machines.

8. A method as in claim 7, wherein the lock is a lock for an object, an asset, or a resource on the first machine.

9. A method as in claim 8, wherein the lock comprises a replicated lock.

10. A method as in claim 8, wherein:

the lock is a lock for an object, an asset, or a resource on the first machine and said first machine comprises a computer;

the updated content comprises an updated value or updated values;

the identifier comprises a name or an identity or an address;

said prior to permitting operation of the lock further comprises prior to permitting acquisition of the lock;

said prior to said second machine acquiring the lock comprises at least one of;

(i) prior to permitting acquisition or operation of the lock;

(ii) prior to permitting acquisition or operation of the same lock on said object, asset, or resource; and

(iii) prior to permitting acquisition or operation of a corresponding replicated lock on an object, asset, or resource.

11. A method as in claim 7, wherein an acquired lock comprises an acquired lock on an object, asset, or resource.

12. A method as in claim 8, wherein an object, asset, or resource is the object, asset, or resource to which said updated content is to be written.

13. A method as in claim 7, wherein said first machine comprises a computer.

14. A method as in claim 7, wherein the updated content comprises an updated value or updated values.

15. A method as in claim 7, wherein the identifier comprises a name or an identity or an address.

16. A method as in claim 7, wherein said prior to said second machine acquiring the lock comprises prior to permitting acquisition or operation of the lock.

17. A method as in claim 16, wherein the prior to permitting operation of the lock further comprises prior to permitting acquisition of the lock.

18. A method as in claim 7, wherein said prior to said second machine acquiring the lock comprises prior to permitting acquisition or operation of the same lock on an object, asset, or resource.

19. A method as in claim 7, wherein said prior to said second machine acquiring the lock comprises prior to permitting acquisition or operation of a corresponding replicated lock on an object, asset, or resource.

20. A method as in claim 1, wherein:

the lock comprises a replicated lock for an object, an asset, or a resource on the first machine;

prior to acquiring the lock permitting acquisition or operation of a corresponding replicated lock on said object, asset, or resource;

the transmitting of the recorded identifier and the updated content to said second machine is performed so that said second machine upon operating on said lock has already received said updated content without waiting for ordinary updating of said written-to replicated memory location.

21. A method as in claim 20, wherein the transmitting of step (d) comprises transmitting only the final content of each written-to memory location of step (b).

22. A method as in claim 20, wherein the transmitting of step (d) comprises transmitting all updated contents of each written-to memory location including a final updated content and an intermediary changed content.

23. A method as in claim 20, wherein in the transmitting of step (d), the prior to permitting operation of the lock by the second machine includes, requiring said second machine to receive the transmitted identifier and content of the written-to replicated memory location transmitted by the first machine, and to have updated content of corresponding identified local replicated memory locations of the second machine with the updated content of the received transmissions of the first machine.

24. A method as in claim 1, wherein the transmitting of step (d) comprises transmitting only the final content of each written-to memory location of step (b).

25. A method as in claim 1, wherein the transmitting of step (d) comprises transmitting all of the updated contents of each written-to memory location including a final updated content and an intermediary changed content.

26. A method as in claim 1, wherein in the transmitting of step (d), the prior to permitting operation of the lock by the second machine includes, requiring said second machine to receive the transmitted identifier and content of the written-to replicated memory location transmitted by the first machine, and to have updated the content of corresponding identified local replicated memory locations of the second machine with the updated content of the received transmissions of the first machine.

27. A method as in claim 1, wherein the transmission of step (d) is transmitted at a higher priority than a lower priority for in due course transmissions where updating includes updating replicated memory locations so that they are updated to remain substantially similar upon occasion of any one of said plurality of machines simultaneously operating an application program causing modification of the content of said replicated memory location of step (a).

28. A method as in claim 1, wherein the transmissions of step (d) and in due course transmissions are a single transmission and not two independent transmissions.

29. A method as in claim 1, wherein the transmissions of step (d) are sent to two or more machines.

30. A method as in claim 29, further including the further step of:

plural machines receiving the transmissions of step (d) and updating identified local replicated memory locations with received corresponding updated content.

31. A method as in claim 1, wherein the local memory of each said plurality of machines are independent of each other machine.

32. A method as in claim 1, each of said processor of each said plurality of machines may only access the local memory of the same machine.

33. A method as in claim 1, wherein at least a first application program written to operate on a single one of said plurality of machines, is operating substantially simultaneously on different ones of said plurality of machines.

34. A method as in claim 1, wherein an application program operating substantially simultaneously on each of said different ones of said plurality of machines may only access the local memory of the same machine.

35. A method as in claim 34, wherein said access is satisfied by the local memory of the same machine independently of and without the aid of the local memories of any other machine.

36. A method as in claim 35, wherein said access includes (or is restricted to) reading and/or writing content resident within said local memory of the same machine.

37. A method as in claim 35, wherein said access includes (or is restricted to) reading and/or writing content of real or virtual memory addresses resident within said local memory of the same machine.

38. A method as in claim 35, wherein said access is restricted to reading and/or writing content resident within said local memory of the same machine.

39. A method as in claim 35, wherein said access is restricted to reading and/or writing content of real or virtual memory addresses resident within said local memory of the same machine.

40. A method as in claim 34, wherein at least one memory location and/or memory content of said application program is substantially similarly replicated in said local memories of said plurality of machines.

41. A method as in claim 40, wherein said at least one substantially similarly replicated memory location and/or memory content are stored non-identically in said local memories of said different ones of said plurality of machines.

42. A method as in claim 41, wherein said at least one substantially similarly replicated memory location and/or memory content are updated in due course to remain substantially similar upon occasion of any one of said plurality of machines simultaneously operating said application program modifying, or causing to be modified, the content of said at least one substantially similarly replicated memory location and/or memory content.

43. A method as in claim 1, wherein:

said local memory of each said plurality of machines are independent of local memory of each other machine;

said processor of each said plurality of machines may only read said local memory of the same machine in which the processor is located;

at least a first application program written to operate on a single one of said plurality of machines, is operating substantially simultaneously on different ones of said plurality of machines;

said application program operating substantially simultaneously on each of said different ones of said plurality of machines, may only access said local memory of the same machine;

said access of said application program is satisfied by said local memory of the same machine independently of or without the aid of said local memory of any other machine;

said access of said application program includes reading and/or writing content or values stored or resident within said local memory of the same machine, or said access of said application program is restricted to reading and/or writing content or values stored or resident within said local memory of the same machine;

said access of said application program includes reading and/or writing content or values of real or virtual memory addresses of or resident within said local memory of the same machine, or said access of said application program is restricted to reading and/or writing content or values of real or virtual memory addresses of or resident within said local memory of the same machine;

at least one memory location and/or memory value of said application program is substantially similarly replicated in said local memory of each of said different ones of said plurality of machines;

said at least one substantially similarly replicated memory location and/or value are stored non-identically in said local memory of each of said plurality of machines;

said substantially similarly replicated memory location and/or memory value are updated through in-due-course updating to remain substantially similar upon occasion of any one of said plurality of machines simultaneously operating said first application program modifying, or causing to be modified, content of said substantially similarly replicated memory location and/or memory value; and

said in-due-course updating provides that said at least one replicated memory location is updated to remain substantially similar upon occasion of any one of said plurality of machines simultaneously operating said application program causing modification of the content of said at least one replicated memory location.

44. An article of manufacture comprising a non-transitory computer readable memory device containing computer-readable instructions therein which, when executed, carry out a method in at least one single computer capable of interoperating with at least one other computer coupled to at least one said single computer at least intermittently via a communications network to form a multiple computer system having a plurality of computers, wherein each computer of the multiple computer system has a local memory and the multiple computer system operating in a replicated shared memory arrangement, the method comprising:

a. acquiring a lock on a first machine having a replicated memory location;

b. recording an identifier of the replicated memory location written-to with an updated content prior to releasing said lock on the first machine;

c. releasing said lock on the first machine; and

d. prior to permitting an operation of the lock by a second machine, transmitting the recorded identifier and the updated content to said second machine.

45. A multiple computer system comprising:

a plurality of computers each including a processor and a local memory coupled with the processor including a first computer and a second computer interconnected via a communications link or network operating in a replicated shared memory arrangement;

means for acquiring a lock on the first computer having a replicated memory location;

means for recording an identifier of the replicated memory location written-to with an updated content prior to releasing said lock on the first computer;

means for releasing said lock on the first computer; and

means for transmitting, prior to permitting an operation of the lock by the second computer, the recorded identifier and the updated content to said second computer.

46. A method carried out in a single computer that operates in a multiple computer system comprising a plurality of computers each having a local processor and a local memory coupled to the local processor, and each of the plurality of computers being interconnected via a communications network, and operating in a replicated shared memory arrangement, the method comprising:

acquiring a lock;

recording an identifier of a replicated memory location written-to with an updated content prior to releasing said lock;

releasing said lock; and

communicating, prior to permitting an operation of the lock, the recorded identifier and the updated content.

47. A single computer intended to operate in a multiple computer system which comprises a plurality of computers each having a local processor and a local memory coupled to the local processor, and each of the plurality of computers being interconnected via a communications network, and operating in a replicated shared memory arrangement, the single computer comprising:

means for acquiring a lock;

means for recording an identifier of a replicated memory location written-to with an updated content prior to releasing said lock;

means for releasing said lock; and

means for communicating, prior to permitting an operation of the lock, the recorded identifier and the updated content.