Method and system for target dependent data dissection and application thereof

1 . A method, comprising:

specifying one of different targets with respect to one of different network management tasks, wherein each of the different network management tasks manages a corresponding one of different operational aspects of a network;

obtaining one of different dissection objective functions (DOFs) with respect to the specified target, wherein the obtained DOF characterizes an operational aspect of the network to facilitate a network management task to manage an operational aspect of the network, and wherein different DOFs correspond to the different targets, respectively;

extracting multiple features of a data set collected from the operation of the network;

generating a plurality of combinations of the multiple features;

dissecting the data set based on each of the plurality of combinations of the multiple features to generate a partition of the network with a set of sub-network segments;

computing, for each sub-network segment in each of the partitions, a DOF score characterizing the operational aspect of the sub-network segment;

identifying, from the sub-network segments in the plurality of partitions, significant sub-network segments based on their respective DOF scores; and

determining, by the network management task, an action to control the operational aspect of each of the most significant sub-network segments.

2 . The method of claim 1 , wherein the multiple features represent each of the sub-network segments in the network and are determined according to the target.

3 . The method of claim 1 , wherein the dissecting the data set based on each of the plurality of combinations comprises:

determining a sequence of groups of the multiple features in each of the plurality combinations to be applied in an order, wherein each of the groups includes some of the multiple features; and

partitioning data in the data set in a hierarchical manner using each of the sequence of groups in the order based on values of the some of the multiple features in the group.

4 . The method of claim 3 , wherein the plurality of combinations represents corresponding plurality of ways to partition the data set based on the multiple features.

5 . The method of claim 3 , wherein the dissecting the data set based on a combination of the multiple features comprises:

applying each of the one or more groups according to the order by,

determining the values of the some of the multiple features in the group based on the data in the data set that has a size satisfying a predetermined criterion,

dividing the data into different sets according to the determined values, and

creating a sub-network segment for each of the different sets; and

generating the partition of the data set based on the sub-network segments.

6 . The method of claim 1 , wherein the computing a DOF score comprises:

obtaining information from the data set relating to the operational aspect of the sub-network; and

determining the DOF score for the sub-network segment based on the determined information that characterizes the operational aspect of the sub-network segment.

7 . The method of claim 1 , wherein the identifying the significant sub-network segments comprises:

accessing a preset criterion provided with respect to the network management task;

evaluating the DOF scores of the sub-network segments in the plurality of partitions against the preset criterion; and

selecting the significant sub-network segments with DOF scores that satisfy the preset criterion.

8 . A machine-readable and non-transitory medium having information recorded thereon, wherein the information, when read by the machine, causes the machine to perform the following steps:

specifying one of different targets with respect to one of different network management tasks, wherein each of the different network management tasks manages a corresponding one of different operational aspects aspect of a network;

obtaining one of different dissection objective functions (DOFs) with respect to the specified target, wherein the obtained DOF characterizes an operational aspect of the network to facilitate a network management task to manage an operational aspect of the network;

extracting multiple features of a data set collected from the operation of the network;

generating a plurality of combinations of the multiple features;

dissecting the data set based on each of the plurality of combinations of the multiple features to generate a partition of the network with a set of sub-network segments;

computing, for each sub-network segment in each of the partitions, a DOF score characterizing the operational aspect of the sub-network segment;

identifying, from the sub-network segments in the plurality of partitions, significant sub-network segments based on their respective DOF scores; and

determining, by the network management task, an action to control the operational aspect of each of the most significant sub-network segments.

9 . The medium of claim 8 , wherein the multiple features represent each of the sub-network segments in the network and are determined according to the target.

10 . The medium of claim 8 , wherein the dissecting the data set based on each of the plurality of combinations comprises:

determining a sequence of groups of the multiple features in each of the plurality combinations to be applied in an order, wherein each of the groups includes some of the multiple features; and

partitioning data in the data set in a hierarchical manner using each of the sequence of groups in the order based on values of the some of the multiple features in the group.

11 . The medium of claim 10 , wherein the plurality of combinations represents corresponding plurality of ways to partition the data set based on the multiple features.

12 . The medium of claim 10 , wherein the dissecting the data set based on a combination of the multiple features comprises:

applying each of the one or more groups according to the order by,

determining the values of the some of the multiple features in the group based on the data in the data set that has a size satisfying a predetermined criterion,

dividing the data into different sets according to the determined values, and

creating a sub-network segment for each of the different sets; and

generating the partition of the data set based on the sub-network segments.

13 . The medium of claim 8 , wherein the computing a DOF score comprises:

obtaining information from the data set relating to the operational aspect of the sub-network; and

determining the DOF score for the sub-network segment based on the determined information that characterizes the operational aspect of the sub-network segment.

14 . The medium of claim 8 , wherein the identifying the significant sub-network segments comprises:

accessing a preset criterion provided with respect to the network management task;

evaluating the DOF scores of the sub-network segments in the plurality of partitions against the preset criterion; and

selecting the significant sub-network segments with DOF scores that satisfy the preset criterion.

15 . A system comprising:

a network management application implemented by a processor and configured for:

specifying one of different targets with respect to one of different network management tasks, wherein each of the different network management tasks manages a corresponding one of different operational aspects of a network, and

obtaining one of different dissection objective functions (DOFs) with respect to the specified target, wherein the obtained DOF characterizes an operational aspect of the network to facilitate a network management task to manage an operational aspect of the network, and wherein different DOFs correspond to the different targets, respectively; and

a targeted hierarchical data dissection (THDD) engine implemented by a processor and configured for:

extracting multiple features of a data set collected from the operation of the network,

generating a plurality of combinations of the multiple features,

dissecting the data set based on each of the plurality of combinations of the multiple features to generate a partition of the network with a set of sub-network segments,

computing, for each sub-network segment in each of the partitions, a DOF score characterizing the operational aspect of the sub-network segment, and

identifying, from the sub-network segments in the plurality of partitions, significant sub-network segments based on their respective DOF scores, wherein

the network management application is further configured for determining, by the network management task, an action to control the operational aspect of each of the most significant sub-network segments.

16 . The system of claim 15 , wherein

the multiple features represent each of the sub-network segments in the network and are determined according to the target; and

the plurality of combinations represents corresponding plurality of ways to partition the data set based on the multiple features.

17 . The system of claim 15 , wherein the dissecting the data set based on each of the plurality of combinations comprises:

determining a sequence of groups of the multiple features in each of the plurality combinations to be applied in an order, wherein each of the groups includes some of the multiple features; and

partitioning data in the data set in a hierarchical manner using each of the sequence of groups in the order based on values of the some of the multiple features in the group.

18 . The system of claim 17 , wherein the dissecting the data set based on a combination of the multiple features comprises:

applying each of the one or more groups according to the order by,

determining the values of the some of the multiple features in the group based on the data in the data set that has a size satisfying a predetermined criterion,

dividing the data into different sets according to the determined values, and

creating a sub-network segment for each of the different sets; and

generating the partition of the data set based on the sub-network segments.

19 . The system of claim 15 , wherein the computing a DOF score comprises:

obtaining information from the data set relating to the operational aspect of the sub-network; and

determining the DOF score for the sub-network segment based on the determined information that characterizes the operational aspect of the sub-network segment.

20 . The system of claim 15 , wherein the identifying the significant sub-network segments comprises:

accessing a preset criterion provided with respect to the network management task;

evaluating the DOF scores of the sub-network segments in the plurality of partitions against the preset criterion; and

selecting the significant sub-network segments with DOF scores that satisfy the preset criterion.