Systems and methods for processing complex data sets

• US 6,915,212 B2
• Filed: 05/08/2003
• Issued: 07/05/2005
• Est. Priority Date: 05/08/2003
• Status: Expired due to Fees

• Current Assignee: Tsunami Development LLC
• Original Assignee: MOAC LLC
• Inventors: Kamps, Bill
• Primary Examiner(s): McElheny, Jr., Donald
• Application Number:
  US10/434,723
• Publication Number:
  US 20040225443A1
• Time in Patent Office: 789 Days
• Field of Search: 702/2, 702/5, 702/14, 702/16, 702/17, 702/18
• US Class Current: 702/14
• CPC Class Codes: G01V 1/28 Processing seismic data, e.g. for interpretation or for event detection G01V1/48 takes precedence
• CPC Class Codes: G01V 2210/57 Trace interpolation or extrapolation, e.g. for virtual receiver; Anti-aliasing for missing receivers

Abstract

Various systems and methods of the present invention provide for distributing access to a dataset to a plurality of processing nodes where the dataset is processed to produce node specific outputs. Distribution can be accomplished by a chain, or star-chain distribution model. Some systems and methods of the present invention provide for check-pointing and restarting improperly terminated processes. Other systems and methods provide for computing a coherent result using a cluster of heterogeneous nodes.

Claims

1. A processing cluster for executing a distributed processing operation on a large dataset, wherein multiple processing platforms perform separate, coordinated processing steps relative to portions of the dataset so as to collectively execute the distributed processing operation, the processing cluster comprising:

a database comprising said dataset;

a first node, associated with a first processing platform, communicably coupled to the database;

a second node, associated with a second processing platform, communicably coupled to the first node, wherein the second node is configured to receive at least a first portion of the dataset from the first node via a first communication channel between said first node and said second node; and

a third node, associated with a third processing platform, communicably coupled to the second node, wherein the third node is configured to receive at least a second portion of the dataset from the second node via a second communication channel between said second node and said third node;

said first, second and third processing platforms thereby being operative for serial transfer of data of said dataset therebetween via said first and second communication channels free from direct transfer of data from said database to either of said second and third nodes;

wherein said first, second and third platforms are operative for executing respective first, second and third separate, coordinated processing steps of said distributed processing operation for said large dataset.

2. The processing cluster of claim 1, wherein the second node is communicably coupled to the first node, and the third node is communicably coupled to the second node via a network switch.

3. The processing cluster of claim 2, wherein the third node is further communicably coupled to the first node via the network switch.

4. The processing cluster of claim 1, wherein the dataset comprises a first data section and a second data section, and wherein the second node is configured to receive the first data section from the first node, the processing cluster further comprising:

a fourth node communicably coupled to the first node, wherein the fourth node is configured to receive the second data section from the first node; and

a fifth node communicably coupled to the fourth node, wherein the fifth node is configured to receive at least a portion of the second data section from the fourth node.

5. The processing cluster of claim 1, wherein the first node comprises a computer readable medium, and wherein the computer readable medium includes:

instructions executable by the first node to determine if the second node and the third node terminate properly.

6. The processing cluster of claim 5, wherein the computer readable medium further includes:

instructions executable by the first node to identify a processing checkpoint of the second node; and

instructions to restart the second node to processing check-point.

7. The processing cluster of claim 1, wherein the first node comprises a computer readable medium, and wherein the computer readable medium includes:

instructions executable by the first node to convert at least a portion of the dataset to a global type.

8. The processing cluster of claim 7, wherein the computer readable medium is a first computer readable medium, wherein the second node comprises a second computer readable medium, and wherein the second computer readable medium includes:

instructions executable by the second node to convert at least a portion of the dataset to a local type.

9. The processing cluster of claim 1, wherein the processing cluster comprises heterogeneous nodes.

10. The processing cluster of claim 9, wherein the first node is a workstation, and the second node is a personal computer.

11. The processing cluster of claim 1, wherein the dataset comprises input seismic trace data, and wherein the second node is configured to create an image of a first region based at least in part on the input seismic trace data, and the third node is configured to create an image of a second region based at least in part on the input seismic trace data.

12. A data processing system, the system comprising:

a database comprising input data;

a master node communicably coupled to a chain of at least two sub-nodes, wherein the chain of at least two sub-nodes includes at least one sub-node configured to receive the input data from a preceding sub-node such that said master node, said one sub-node and said preceding sub-node are operative for serial transfer of said input data therebetween free from direct transfer of said input data between said master node and said at least one sub-node;

wherein said preceding sub-node and said at least one sub-node are operative for executing respective first and second separate, coordinated processing steps of a distributed processing operation with respect to said input data.

13. The system of claim 12, wherein the chain of at least two sub-nodes is a first chain, and wherein the master node is communicably coupled lo a second chain of at least two sub-nodes.

14. A seismic data processing cluster for executing a distributed processing operation on a large dataset, wherein multiple processing platforms perform separate, coordinated processing steps relative to portions of the dataset so as to collectively execute the distributed processing operation, the processing cluster comprising:

a database comprising said dataset of input seismic trace data;

a first node, associated with a first processing platform, communicably coupled to the database;

a second node, associated with a second processing platform, communicably coupled to the first node, wherein the second node is configured to receive at least a first portion of the dataset from the first node via a first communication channel between said first node and said second node; and

a third node, associated with a third processing platform, communicably coupled to the second node, wherein the third node is configured to receive at least a second portion of the dataset from the second node via a second communication channel between said second node and said third node;

said first, second and third processing platforms thereby being operative for serial transfer of data of said dataset therebetween via said first and second communication channels free from direct transfer of data from said database to either of said second and third nodes;

wherein said first, second and third platforms are operative for executing respective first, second and third separate, coordinated processing steps of said distributed processing operation for said large dataset.

15. The seismic data processing cluster of claim 14, wherein the second node is communicably coupled to the first node, and the third node is communicably coupled to the second node via a network switch.

16. The seismic data processing cluster of claim 15, wherein the third node is further communicably coupled to the first node via the network switch.

17. The seismic data processing cluster of claim 14, wherein the dataset comprises a first data section and a second data section, and wherein the second node is configured to receive the first data section from the first node, the processing cluster further comprising:

a fourth node communicably coupled to the first node, wherein the fourth node is configured to receive the second data section from the first node; and

a fifth node communicably coupled to the fourth node, wherein the fifth node is configured to receive at least a portion of the second data section from the fourth node.

18. The seismic data processing cluster of claim 14, wherein the first node comprises a computer readable medium, and wherein the computer readable medium includes:

instructions executable by the first node to determine if the second node and the third node terminate properly.

19. The seismic data processing cluster of claim 18, wherein the computer readable medium further includes:

instructions executable by the first node to identify a processing checkpoint of the second node; and

instructions to restart the second node at the processing check-point.

20. The seismic data processing cluster of claim 14, wherein the first node comprises a computer readable medium, and wherein the computer readable medium includes:

instructions executable by the first node to convert at least a portion of the dataset to a global type.

21. The seismic data processing cluster of claim 20, wherein the computer readable medium is a first computer readable medium, wherein the second node comprises a second computer readable medium, and wherein the second computer readable medium includes:

instructions executable by the second node to convert at least a portion of the dataset to a local type.

22. The seismic data processing cluster of claim 14, wherein the processing cluster comprises heterogeneous nodes.

23. The seismic data processing cluster of claim 22, wherein the first node is a workstation, and the second node is a personal computer.

24. The seismic data processing cluster of claim 14, wherein the second node is configured to create an image of a first region based at least in part on the input seismic trace data, and the third node is configured to create an image of a second region based at least in part on the input seismic trace data.

25. A seismic data imaging system, the system comprising:

a database comprising input seismic trace data;

a master node communicably coupled to a chain of at least two sub-nodes, wherein the chain of at least two sub-nodes includes at least one sub-node configured to receive the input seismic trace data from a preceding sub-node such that said master node, said one sub-node and said preceding sub-node are operative for serial transfer of said input data therebetween free from direct transfer of said input data between said master node and said at least one sub-node;

wherein said preceding sub-node and said at least one sub-node are operative for executing respective first and second separate, coordinated processing steps of a distributed processing operation with respect so said input data.

26. The system of claim 25, wherein the chain of at least two sub-nodes is a first chain, and wherein the master node is communicably coupled to a second chain of at least two sub-nodes.

27. A computer readable medium, the computer readable medium comprising computer executable instructions to:

receive input seismic data from an upstream node, wherein the upstream node is one of a chain of nodes serially communicably coupled to a master node such that said data is obtained from said upstream node free from direct transfer of data from said master node independent of said upstream node;

compute an image of a physical location based at least in part on the input seismic data; and

provide a first output, wherein said first output is combined with a second output of said upstream node to yield a composite result based on a distributed processing operation performed on said input seismic data.

28. The computer readable medium of claim 27, wherein the computer readable medium further comprises computer executable instructions to:

monitor a termination characteristic of a downstream node, wherein the downstream node is one of the chain of nodes.

29. The computer readable medium of claim 28, wherein the compute, readable medium further comprises computer executable instructions to:

identify a check-point of the downstream node; and

restart the downstream node at the check-point.

30. The computer readable medium of claim 27, wherein the computer readable medium further comprises computer executable instructions to:

convert the input seismic data to a global type.

31. The computer readable medium of claim 27, wherein the computer, readable medium further comprises computer executable instructions to:

convert the input seismic data to a local type.

32. The computer readable medium of claim 27, wherein the computer readable medium further comprises computer executable instructions to:

receive job parameters; and

transfer the image to the master node.

33. A computer readable medium, the computer readable medium comprising computer executable instructions to:

access an output trace file, wherein the output trace file identifies a plurality of output seismic traces to be computed;

access a node file, wherein the node file includes an attribute about each node in a chain of nodes; and

assign each of the plurality of output seismic traces to a node in the chain of nodes, wherein computation of each of the plurality of output seismic traces on the assigned nodes completes within a balanced time.

34. The computer readable medium of claim 33, the computer readable medium further comprising computer executable instructions to:

determine a time required to compute an output trace selected from the plurality of output traces on a node selected from the chain of nodes.

35. The computer readable medium of claim 33, the computer readable medium further comprising computer executable instructions to:

monitor computation of an output trace on an assigned node.

36. The computer readable medium of claim 35, the computer readable medium further comprising computer executable instructions to:

identify a check-point of the assigned node; and

restart the assigned node at the check-point.

37. The computer readable medium of claim 33, the computer readable medium further comprising computer executable instructions to:

receive the computed output traces from the respective assigned nodes.

38. The computer readable medium of claim 33, wherein the plurality of output traces is a first plurality of output traces, wherein the chain of nodes is a first chain of nodes, and wherein the first plurality of output traces is a subset of a group of output traces, the computer readable medium further comprising computer executable instructions to:

segregate the group of output traces into at least the first plurality of output traces and a second plurality of output traces;

access an output trace file, wherein the output trace file identifies a plurality of output traces to be computed;

access a node file, wherein the node file includes an attribute about each node in a chain of nodes; and

assign each of the second plurality of output traces to a node in a second chain of nodes.

39. A method of computing, the method comprising:

accessing a dataset;

designating a plurality of nodes as a node chain, wherein the node chain comprises a first node, a second node, and a third node;

serially transferring the dataset from the first node to the second node, and from the second node to the third node such that said third node receives said dataset substantially free from direct communication between said first node and said third node;

processing the dataset on the first node to create a first output, on the second node to create a second output, and on the third node to create a third output; and

assembling the first, second and third outputs to form a coherent output.

40. The method of claim 39, wherein the coherent output is an image of a geologic structure.

41. A method for use in controlling a distributed processing operation for processing seismic data to yield geologic information regarding a subterranean geologic formation, said distributed processing operation involving execution, on multiple platforms, of separated, coordinated processing steps with respect to a common processing job, said method comprising the steps of:

monitoring said distributed processing operation for processing seismic data to identify a malfunction;

identifying a portion of said processing job affected by said malfunction; and

automatically re-tasking at least one processing platform substantially free from any concurrent prompts by a human operator related to said re-tasking, so as to complete said portion of said processing job.

42. A method as set forth in claim 41, wherein said step of monitoring comprises identifying a lack of progress towards completion of said portion of said job.

43. A method as set forth in claim 41, wherein said step of identifying comprises identifying at least one malfunctioning processing platform associated with said malfunction.

44. A method as set forth in claim 43, wherein said step of re-tasking comprises designating at least one processing platform other than said malfunctioning platform to execute said portion of said job.

45. A method as set forth in claim 41, wherein said step of re-tasking comprises automatically restarting said portion of said job substantially free from any concurrent prompt from a human operator related to said restarting.

46. A method as set forth in claim 45, wherein said job is divided into a number of sections separated by check-points, and said step of restarting comprises initiating execution of said portion of said job at a last one of said check-points successfully reached.

47. A method as set forth in claim 41, wherein said distributed processing operation involves a serial data transfer wherein each of a number of processing platforms is assigned to receive data from a designated upstream node and to transmit data to a designated downstream node, and said step of re-tasking comprises changing a designation of one of an upstream node and a downstream node for at least one of said number of processing platforms.

48. A method for use in implementing a distributed processing operation for processing seismic data to yield geologic information regarding a subterranean geologic formation, wherein the seismic data includes data corresponding to a number of traces where each trace reflects a seismic signal received at a sensor location, said method comprising the steps of:

providing a cluster of processing platforms for executing said distributed processing operation, wherein said processing platforms execute separate, coordinated processing steps so as to collectively yield said geologic information;

establishing a first data type for a transfer of data from a first processing platform to a second processing platform of said cluster of processing platforms, wherein said first data type relates to the way that data is represented within a content of said transfer; and

operating said second processing platform to convert said content of said transfer from said first data type to a second data type different than said first data type;

wherein said seismic data is processed in a distributed processing environment involving heterogeneous processing platforms.

49. A method as set forth in claim 48, further comprising the steps of running a first operating system on said first processing platform and running a second operating system, different than said first operating system, on said second processing platform.

50. A method as set forth in claim 48, wherein said first data type reflects a big endian format and said second data type reflects a little endian format.

51. An apparatus for use in controlling a distributed processing operation for processing seismic data to yield geologic information regarding a subterranean geologic formation, said distributed processing operation involving execution, on multiple platforms, of separated, coordinated processing steps with respect to a common processing job, said apparatus comprising the steps of:

a monitoring module for monitoring said distributed processing operation for processing seismic data to identify a malfunction and a portion of said processing job affected by said malfunction; and

a re-tasking module for automatically re-tasking at least one processing platform of said multiple processing platforms substantially free from any concurrent prompts by a human operator related to said re-tasking, so as to complete said portion of said processing job.

52. An apparatus as set forth in claim 51, wherein said monitoring module is further operative for identifying a lack of progress towards completion of said portion of said job.

53. An apparatus as set forth in claim 51, wherein said monitoring module is further operative for identifying at least one malfunctioning processing platform associated with said malfunction.

54. An apparatus as set forth in claim 53, wherein said re-tasking module is operative for designating at least one processing platform other than said malfunctioning platform to execute said portion of said job.

55. An apparatus as set forth in claim 51, wherein said re-tasking module is further operative for automatically restarting said portion of said job substantially free from any concurrent prompt from a human operator related to said starting.

56. An apparatus as set forth in claim 55, wherein said job is divided into a number of sections separated by check-points, and said restarting module is operative for initiating execution of said portion of said job at a last one of said check-points successfully reached.

57. An apparatus as set forth in claim 51, wherein said distributed processing operation involves a serial data transfer wherein each of a number of processing platforms is assigned to receive data from a designated upstream node and to transmit data to a designated downstream node, and said re-tasking module is operative for changing a designation of one of an upstream node and a downstream node for at least one of said number of processing platforms.

58. An apparatus for use implementing a distributed processing operation for processing seismic data to yield geologic information regarding a subterranean geologic formation, wherein the seismic data includes data corresponding to a number of traces where each trace reflects a seismic signal received at a sensor location, said apparatus comprising:

a cluster of processing platforms configured for executing said distributed processing operation, wherein said processing platforms execute separate, coordinated processing steps so as to collectively yield said geologic information, wherein said cluster is configured to establish a first data type for transfer of data from a first processing platform to a second processing platform of said cluster of processing platforms, said first data type relating to the way that data is represented within a content of said transfer;

said second processing platform being operative to convert said content of said transfer from said first data type to a second data type different than said first data type, wherein said seismic data is processed in a distributed processing environment involving heterogeneous platforms.

59. An apparatus as set forth in claim 58, wherein said first processing platform is operative for running a first operating system and said second processing platform is operative for running a second operating system different than said first operating system.

60. An apparatus as set forth in claim 58, wherein said first data type reflects a big endian format and said second data type reflects a little endian format.

61. A processing cluster for executing a distributed processing operation on a large dataset of seismic data, wherein multiple processing platforms perform separate, coordinated processing steps relative to portions of the dataset so as to collectively execute the distributed processing operation, the processing cluster comprising:

a database comprising said dataset of seismic data;

a first node, associated with a first processing platform, communicably coupled to the database;

a second node, associated with a second processing platform, communicably coupled to the first node, wherein the second node is configured to receive at least a first portion of the dataset of seismic data from the first node via a first communication channel between said first node and said second node; and

a third node, associated with a third processing platform, communicably coupled to the second node, wherein the third node is configured to receive at least a second portion of the dataset of seismic data from the second node via a second communication channel between said second node and said third node;

said first, second and third processing platforms thereby being operative for serial transfer of data of said dataset of seismic data therebetween via said first and second communication channels free from direct transfer of data from said database to either of said second and third nodes;

wherein said first, second and third platforms are operative for executing respective first, second and third separate, coordinated processing steps of said distributed processing operation for said dataset of seismic data;

wherein said first node and second node are configured for transfer of data of a first data type therebetween, said first data type relating to the way that data is represented within a content of said transfer; and

said second node is operative to convert said content of said transfer from said first data type to a second data type different than said first data type, wherein said seismic data is processed in a distributed processing environment involving heterogeneous platforms.

62. A processing cluster for executing a distributed processing operation on a large dataset of seismic data, wherein multiple processing platforms perform separate, coordinated processing steps relative to portions of the dataset so as to collectively execute the distributed processing operation, the processing cluster comprising:

a database comprising said dataset of seismic data;

a first node, associated with a first processing platform, communicably coupled to the database;

a second node, associated with a second processing platform, communicably coupled to the first node, wherein the second node is configured to receive at least a first portion of the dataset of seismic data from the first node via a first communication channel between said first node and said second node; and

a third node, associated with a third processing platform, communicably coupled to the second node, wherein the third node is configured to receive at least a second portion of the dataset of seismic data from the second node via a second communication channel between said second node and said third node;

said first, second and third processing platforms thereby being operative for serial transfer of data of said dataset of seismic data therebetween via said first and second communication channels free from direct transfer of data from said database to either of said second and third nodes;

wherein said first, second and third platforms are operative for executing respective first, second and third separate, coordinated processing steps of said distributed processing operation for said dataset of seismic data;

a monitoring module for monitoring said distributed processing operation to identify a malfunction and one of said first, second and third nodes affected by said malfunction; and

a re-tasking module for automatically re-tasking one of said first, second and third nodes substantially free from any concurrent prompts by a human operator related to said re-tasking.