COMMUNICATION NETWORK FAILURE DETECTION SYSTEM, AND COMMUNICATION NETWORK FAILURE DETECTION METHOD AND FAILURE DETECTION PROGRAM

• US 20090167520A1
• Filed: 07/11/2006
• Published: 07/02/2009
• Est. Priority Date: 07/11/2005
• Status: Active Grant

• Current Assignee: Cloud BYTE LLC
• Original Assignee: NEC Corporation
• Inventors: Watanabe, Yoshinori, Matsunaga, Yasuhiko
• Granted Patent: US 8,065,568 B2
• Time in Patent Office: Days
• US Class Current: 340/507
• CPC Class Codes: H04L 41/064 involving time analysis
• CPC Class Codes: H04L 41/065 involving logical or physical relationship, e.g. grouping and hierarchies
• CPC Class Codes: H04L 43/00 Arrangements for monitoring or testing data switching networks
• CPC Class Codes: H04L 43/022 by sampling
• CPC Class Codes: H04L 43/16 Threshold monitoring

Abstract

The system for detecting a failure on a communication network according to an objective basis by analyzing an alarm log output by a management server of the communication network and evaluating a degree of abnormality of the communication network tracing back to a more significant factor causing the alarm, calculates an occurrence intensity of the more significant factor 500 causing an alarm based on the recording contents of the alarm, and detects a failure derived from the more significant factor of the communication network based on the calculated intensity of occurrence of said more significant factor 500.

Claims

1-33. -33. (canceled)

34. In a communication network for which contents of an alarm issued at the time of a failure of communication are recorded, a communication network failure detection system, comprising:

an occurrence intensity calculating unit for a more significant factor for calculating an occurrence intensity of said more significant factor causing an alarm by analyzing the recording contents of the alarm;

an occurrence intensity probability distribution calculating unit for calculating a probability distribution of an occurrence intensity of said more significant factor in said communication network at a normal state;

a degree of abnormality calculating unit for calculating, with respect to a calculated occurrence intensity of said more significant factor, a degree of probability deviation from an occurrence intensity probability distribution of said more significant factor at said normal state as a degree of abnormality; and

a failure detecting unit for detecting a failure derived from said more significant factor in said communication network by comparing said degree of abnormality and a threshold value applied in advance.

35. The communication network failure detection system according to claim 34, wherein said occurrence intensity calculating unit for said more significant factor includes a time variation parameter extracting unit for extracting a parameter indicative of a time variation in an alarm occurrence frequency from the recording contents of said alarm to consider a value of said parameter indicative of a time variation as an occurrence intensity of said more significant factor.

36. The communication network failure detection system according to claim 34, wherein said occurrence intensity probability distribution calculating unit calculates a probability distribution of an occurrence intensity of said more significant factor in said communication network at the normal state based on an estimation result obtained by the occurrence intensity calculating unit for the more significant factor during a past period when a communication device having issued the alarm operated normally.

37. The communication network failure detection system according to claim 34, wherein said occurrence intensity calculating unit for the more significant factor includes a space variation parameter extracting unit for extracting a parameter indicative of a space variation in an alarm occurrence frequency from the recording contents of said alarm to consider a value of said parameter indicative of a space variation as an occurrence intensity of said more significant factor.

38. The communication network failure detection system according to claim 34, wherein said occurrence intensity probability distribution calculating unit calculates a probability distribution of an occurrence intensity of said more significant factor in the communication network at the normal state based on an estimation result obtained by the occurrence intensity calculating unit for the more significant factor in other communication device geographically adjacent to a communication device having issued the alarm.

39. A communication network failure detection system, comprising:

an occurrence intensity calculating unit for a more significant factor for calculating an occurrence intensity of a more significant factor causing an alarm by analyzing the recording contents of said alarm;

an occurrence intensity probability distribution calculating unit for calculating a probability distribution of an occurrence intensity of said more significant factor in a communication network at a normal state;

a degree of abnormality calculating unit for calculating, with respect to a calculated occurrence intensity of said more significant factor, a degree of probability deviation from an occurrence intensity probability distribution of said more significant factor at the normal state as a degree of abnormality; and

a failure detecting unit for detecting a failure of the communication network by comparing said degree of abnormality and a threshold value applied in advance, wherein

said occurrence intensity calculating unit for the more significant factor includes a time variation parameter extracting unit for extracting a parameter indicative of a time variation in an alarm occurrence frequency from the recording contents of said alarm, and a space variation parameter extracting unit for extracting a parameter indicative of a space variation in an alarm occurrence frequency to consider values of said parameter indicative of a time variation and said parameter indicative of a space variation as an occurrence intensity of said more significant factor.

40. The communication network failure detection system according to claim 39, wherein said occurrence intensity probability distribution calculating unit calculates a probability distribution of an occurrence intensity of said more significant factor in the communication network at the normal state based on an estimation result obtained by said occurrence intensity calculating unit for the more significant factor during a past period when a communication device having issued the alarm operated normally and an estimation result obtained by said occurrence intensity calculating unit for the more significant factor in other communication device geographically adjacent to the communication device having issued the alarm.

41. The communication network failure detection system according to claim 39, wherein said time variation parameter extracting unit and said space variation parameter extracting unit obtain statistically independent time variation component and space variation component from an alarm occurrence frequency by using an independent component analysis method to consider the components as the parameter indicative of a time variation and the parameter indicative of a space variation, respectively.

42. The communication network failure detection system according to claim 39, wherein with an upper side probability or a lower side probability or a both-side probability of the occurrence intensity probability distribution as a degree of abnormality, individually with respect to an occurrence intensity of said more significant factor obtained from said parameter indicative of a time variation and an occurrence intensity of said more significant factor obtained from said parameter indicative of a space variation, said degree of abnormality calculating unit multiplies occurrence intensities of a plurality of more significant factors by the respective degrees of abnormality obtained to consider a result as a degree of abnormality as a whole.

43. The communication network failure detection system according to claim 39, wherein said failure detecting unit detects a failure by using both said degree of abnormality obtained from said parameter indicative of a time variation and said degree of abnormality obtained from said parameter indicative of a space variation.

44. In a communication network for which contents of an alarm issued at the time of a failure of communication are recorded, a communication network failure detecting method, comprising the steps of:

a occurrence intensity calculating step for a more significant factor of calculating an occurrence intensity of said more significant factor causing an alarm by analyzing the recording contents of the alarm;

a occurrence intensity probability distribution calculating step of calculating a probability distribution of an occurrence intensity of said more significant factor in said communication network at a normal state;

a degree of abnormality calculating step of calculating, with respect to a calculated occurrence intensity of said more significant factor, a degree of probability deviation from an occurrence intensity probability distribution of said more significant factor at said normal state as a degree of abnormality; and

a failure detecting step of detecting a failure derived from said more significant factor in said communication network by comparing said degree of abnormality and a threshold value applied in advance.

45. The communication network failure detecting method according to claim 44, wherein said occurrence intensity calculating step for said more significant factor includes a time variation parameter extracting step of extracting a parameter indicative of a time variation in an alarm occurrence frequency from the recording contents of said alarm to consider a value of said parameter indicative of a time variation as an occurrence intensity of said more significant factor.

46. The communication network failure detecting method according to claim 44, wherein at said occurrence intensity probability distribution calculating step, a probability distribution of an occurrence intensity of said more significant factor in said communication network at the normal state is calculated based on an estimation result obtained by the occurrence intensity calculating step for the more significant factor during a past period when a communication device having issued the alarm operated normally.

47. The communication network failure detecting method according to claim 44, wherein said occurrence intensity calculating step for the more significant factor includes a space variation parameter extracting step of extracting a parameter indicative of a space variation in an alarm occurrence frequency from the recording contents of said alarm to consider a value of said parameter indicative of a space variation as an occurrence intensity of said more significant factor.

48. The communication network failure detecting method according to claim 44, wherein at said occurrence intensity probability distribution calculating step, a probability distribution of an occurrence intensity of said more significant factor in the communication network at the normal state is calculated based on an estimation result obtained by the occurrence intensity calculating step for the more significant factor in other communication device geographically adjacent to a communication device having issued the alarm.

49. A communication network failure detecting method, comprising the steps of:

a occurrence intensity calculating step for a more significant factor of calculating an occurrence intensity of said more significant factor causing an alarm by analyzing the recording contents of said alarm;

a occurrence intensity probability distribution calculating step of calculating a probability distribution of an occurrence intensity of said more significant factor in a communication network at a normal state;

a degree of abnormality calculating step of calculating, with respect to a calculated occurrence intensity of said more significant factor, a degree of probability deviation from an occurrence intensity probability distribution of said more significant factor at the normal state as a degree of abnormality; and

a failure detecting step of detecting a failure of the communication network by comparing said degree of abnormality and a threshold value applied in advance, wherein

said occurrence intensity calculating step for the more significant factor

includes a time variation parameter extracting step of extracting a parameter indicative of a time variation in an alarm occurrence frequency from the recording contents of said alarm and the space variation parameter extracting step of extracting a parameter indicative of a space variation in an alarm occurrence frequency, and

considers values of said parameter indicative of a time variation and said parameter indicative of a space variation as an occurrence intensity of said more significant factor.

50. The communication network failure detecting method according to claim 49, wherein at said occurrence intensity probability distribution calculating step, a probability distribution of an occurrence intensity of said more significant factor in the communication network at the normal state is calculated based on an estimation result obtained by said occurrence intensity calculating step for the more significant factor during a past period when a communication device having issued the alarm operated normally and an estimation result obtained by said occurrence intensity calculating step for the more significant factor in other communication device geographically adjacent to the communication device having issued the alarm.

51. The communication network failure detecting method according to claim 49, wherein at said time variation parameter extracting step and said space variation parameter extracting step, statistically independent time variation component and space variation component are obtained from an alarm occurrence frequency by using an independent component analysis method to consider the components as the parameter indicative of a time variation and the parameter indicative of a space variation, respectively.

52. The communication network failure detecting method according to claim 49, wherein at said degree of abnormality calculating step, with an upper side probability or a lower side probability or a both-side probability of the occurrence intensity probability distribution as a degree of abnormality, individually with respect to an occurrence intensity of said more significant factor obtained from said parameter indicative of a time variation and an occurrence intensity of said more significant factor obtained from said parameter indicative of a space variation, occurrence intensities of a plurality of more significant factors are multiplied by the respective degrees of abnormality obtained to consider a result as a degree of abnormality as a whole.

53. The communication network failure detecting method according to claim 49, wherein at said failure detecting step, a failure is detected by using both said degree of abnormality obtained from said parameter indicative of a time variation and said degree of abnormality obtained from said parameter indicative of a space variation.

54. A failure detection program operable in a communication device on a communication network for which contents of an alarm issued at the time of a failure of communication are recorded, which causes said communication device to execute the functions of:

a occurrence intensity calculating function for a more significant factor of calculating an occurrence intensity of said more significant factor causing an alarm by analyzing the recording contents of the alarm;

a occurrence intensity probability distribution calculating function of calculating a probability distribution of an occurrence intensity of said more significant factor in said communication network at a normal state;

a degree of abnormality calculating function of calculating, with respect to a calculated occurrence intensity of said more significant factor, a degree of probability deviation from an occurrence intensity probability distribution of said more significant factor at said normal state as a degree of abnormality; and

a failure detecting function of detecting a failure derived from said more significant factor in said communication network by comparing said degree of abnormality and a threshold value applied in advance.

55. The failure detection program according to claim 54, wherein said function of calculating an occurrence intensity of said more significant factor includes the time variation parameter extracting function of extracting a parameter indicative of a time variation in an alarm occurrence frequency from the recording contents of said alarm to consider a value of said parameter indicative of a time variation as an occurrence intensity of said more significant factor.

56. The failure detection program according to claim 54, wherein as said occurrence intensity probability distribution calculating function, a probability distribution of an occurrence intensity of said more significant factor in said communication network at the normal state is calculated based on an estimation result obtained by the occurrence intensity calculating function for the more significant factor during a past period when a communication device having issued the alarm operated normally.

57. The failure detection program according to claim 54, wherein said function of calculating an occurrence intensity of the more significant factor includes the space variation parameter extracting function of extracting a parameter indicative of a space variation in an alarm occurrence frequency from the recording contents of said alarm to consider a value of said parameter indicative of a space variation as an occurrence intensity of said more significant factor.

58. The failure detection program according to claim 54, wherein as said occurrence intensity probability distribution calculating function, a probability distribution of an occurrence intensity of said more significant factor in the communication network at the normal state is calculated based on an estimation result obtained by the occurrence intensity calculating function for the more significant factor in other communication device geographically adjacent to a communication device having issued the alarm.

59. A failure detection program, which causes a communication device on a communication network to execute the functions of:

a occurrence intensity calculating function for a more significant factor of calculating an occurrence intensity of a more significant factor causing an alarm by analyzing the recording contents of the alarm;

a occurrence intensity probability distribution calculating function of calculating a probability distribution of an occurrence intensity of said more significant factor in the communication network at a normal state;

a degree of abnormality calculating function of calculating, with respect to a calculated occurrence intensity of said more significant factor, a degree of probability deviation from an occurrence intensity probability distribution of said more significant factor at the normal state as a degree of abnormality; and

a failure detecting function of detecting a failure of the communication network by comparing said degree of abnormality and a threshold value applied in advance, wherein

said occurrence intensity calculating function for said more significant factor

includes the time variation parameter extracting function of extracting a parameter indicative of a time variation in an alarm occurrence frequency from the recording contents of said alarm, and the space variation parameter extracting function of extracting a parameter indicative of a space variation in an alarm occurrence frequency, and

considers values of said parameter indicative of a time variation and said parameter indicative of a space variation as an occurrence intensity of said more significant factor.

60. The failure detection program according to claim 59, wherein in said occurrence intensity probability distribution calculating function, a probability distribution of an occurrence intensity of said more significant factor in the communication network at the normal state is calculated based on an estimation result obtained by said occurrence intensity calculating function for the more significant factor during a past period when a communication device having issued the alarm operated normally and an estimation result obtained by said occurrence intensity calculating function for the more significant factor in other communication device geographically adjacent to the communication device having issued the alarm.

61. The failure detection program according to claim 59, wherein in said time variation parameter extracting function and said space variation parameter extracting function, statistically independent time variation component and space variation component are obtained from an alarm occurrence frequency by using an independent component analysis method to consider the components as the parameter indicative of a time variation and the parameter indicative of a space variation, respectively.

62. The failure detection program according to claim 59, wherein in said degree of abnormality calculating function, with an upper side probability or a lower side probability or a both-side probability of the occurrence intensity probability distribution as a degree of abnormality, individually with respect to an occurrence intensity of said more significant factor obtained from said parameter indicative of a time variation and an occurrence intensity of said more significant factor obtained from said parameter indicative of a space variation, occurrence intensities of a plurality of more significant factors are multiplied by the respective degrees of abnormality obtained to consider a result as a degree of abnormality as a whole.

63. The failure detection program according to claim 59, wherein in said failure detecting function, a failure is detected by using both said degree of abnormality obtained from said parameter indicative of a time variation and said degree of abnormality obtained from said parameter indicative of a space variation.