Feature extraction method and apparatus
AssigneeFujitsu Limited
InventorNaoki Akaboshi
Life of this patent
Estimating the term…
Abstract
This method includes: identifying, for each child log having a job type identifier in a message to a lower-layer server and a start time and end time of a processing executed in the lower-layer server, an inclusive parent log that is a parent log having a start time and end time of a processing executed in an upper-layer Web server, wherein the start time and end time of the inclusive parent log define an interval containing a period from the start time to the end time of the child log, and registering the job type identifier in the child log and a path of URI in the inclusive parent log, into a table, and registering a frequency for each combination of the identifier and the path of URI into the table. By using this table, features in the path of URI are extracted for each job type identifier.
Claims6
A computer-readable storage medium storing a program for execution by a computer to perform a feature extraction process for extracting a feature portion of a Universal Resource Identifier (URI), which distinguishes a job executed in a lower-layer server in a multi-layer server system and is included in a message to an upper-layer Web server, said feature extraction process comprising:
- identifying, for each of child logs stored in a child log data storage storing child logs, each having a job type identifier included in a first message to said lower-layer server and a start time and an end time of a processing executed in said lower-layer server in response to said first message, an inclusive parent log that is a parent log having said start time and said end time, which define a time interval containing a time period from said start time to said end time of said child log, from a parent log data storage storing said parent logs, each having said URI included in a second message to said upper-layer Web server and a start time and an end time of a processing executed in said upper-layer Web server in response to the second message, and registering said job type identifier included in said child log and a path of said URI, which is included in said inclusive parent log, into a child-parent relationship candidate table, and counting and registering frequencies for each combination of said job type identifier and said path of said URI into said child-parent relationship candidate table;
- extracting, from said child-parent relationship candidate table, said path of said URI, which is included in said combination whose counted frequency is the maximum, for each of said job type identifiers, and registering a set of said job type identifier and the extracted path of said URI into a child-parent
- identifying, in said child-parent relationship list, said set satisfying a condition that said path of said URI in said set does not match said paths of said URIs in the other sets.
A feature extraction method for extracting a feature portion of a Universal Resource Identifier (URI), which distinguishes a job executed in a lower-layer server in a multi-layer server system and is included in a message to an upper-layer Web server, said feature extraction method comprising:
- providing a child log data storage storing child logs, each having a job type identifier included in a first message to said lower-layer server and a start time and an end time of a processing executed in said lower-layer server in response to said first message, and a parent log data storage storing parent logs, each having said URI included in a second message to said upper-layer Web server and a start time and an end time of a processing executed in said upper-layer Web server in response to the second message;
- identifying, for each of said child logs, an inclusive parent log that is a parent log having said start time and said end time, which define a time interval containing a time period from said start time to said end time of said child log, from said parent log data storage, and registering said job type identifier included in said child log and a path of said URI, which is included in said inclusive parent log, into a child-parent relationship candidate table, and counting and registering frequencies for each combination of said job type identifier and said path of said URI into said child-parent relationship candidate table;
- extracting, from said child-parent relationship candidate table, said path of said URI, which is included in said combination whose counted frequency is the maximum, for each of said job type identifiers, and registering a set of said job type identifier and the extracted path of said URI into a child-parent relationship list;
- and
- identifying, in said child-parent list, said set satisfying a condition that said path of said URI in said set does not match said paths of said URIs in the other sets.
A feature extraction apparatus for extracting a feature portion of a Universal Resource Identifier (URI), which distinguishes a job executed in a lower-layer server in a multi-layer server system and is included in a message to an upper-layer Web server, said feature extraction apparatus comprising:
- a child log data storage device storing child logs, each having a job type identifier included in a first message to said lower-layer server and a start time and an end time of a processing executed in said lower-layer server in response to said first message;
- a parent log data storage device storing parent logs, each having said URI included in a second message to said upper-layer Web server and a start time and an end time of a processing executed in said upper-layer Web server in response to the second message;
- a child-parent relation constructor to identify, for each of said child logs, an inclusive parent log that is a parent log having said start time and said end time, which define a time interval containing a time period from said start time to said end time of said child log, from said parent log data storage, and to register said job type identifier included in said child log and a path of said URI, which is included in said inclusive parent log, into a child-parent relationship candidate table, and to count and register frequencies for each combination of said job type identifier and said path of said URI into said child-parent relationship candidate table, and to extract, from said child-parent relationship candidate table, said path of said URI, which is included in said combination whose counted frequency is the maximum, for each of said job type identifiers, and to register a set of said job type identifier and the extracted path of said URI into a child-parent relationship list;
- and
- a feature portion extractor to identify, in said child-parent list, said set whose path of said URI does not match said paths of said URIs in the other sets exists.
Dependencies are inferred from the claim text.