Error detection/correction and fault detection/recovery – Data processing system error or fault handling – Reliability and availability
Reexamination Certificate
1998-07-01
2001-07-24
Homere, Jean R. (Department: 2777)
Error detection/correction and fault detection/recovery
Data processing system error or fault handling
Reliability and availability
C707S793000
Reexamination Certificate
active
06266788
ABSTRACT:
BACKGROUND OF THE INVENTION
This invention relates generally to a system and method for processing one or more data sets received from one or more computer-based systems and in particular to a system and method for automatically categorizing and characterizing the data sets generated by the computer-based system.
The tremendous expansion of the Internet has led to the expansion of the number features in software applications. The expansion of the Internet has also necessitated the more rapid development of various software applications and has changed various software development methodologies. For example, the technique for beta testing software applications has drastically changed. Prior to the proliferation of the Internet, a company might beta test a software application by distributing the software application by floppy disk to a limited number of beta testers. With the Internet, the process of beta testing requires only that the developer place the beta software application on its web site and then anyone interested in beta testing the software application may do so with almost no expense to the developer. The problem with both of these beta test distribution techniques, however, is that it is difficult for the developer of the software to obtain good feedback from the beta testers.
Therefore, in order to properly beta test a software application and for the developer to benefit from the beta test, it is desirable to provide some medium for the beta testers to communicate with the developer who can gather the beta testers' bug reports and comments and correct any bugs. This process was typically accomplished by a beta test coordinator who was responsible for gathering the relevant information and routing the bug reports to the appropriate engineers. It is desirable to provide a system that automatically retrieves the bug reports and comments from beta testers.
In order to provide quality assurance (QA) feedback to a user of the software application, it is additionally desirable to be able to recreate a user's problem so that the Quality Assurance person can quickly help the user. In some conventional systems, the Quality Assurance person attempts to recreate the problem based on a user's recollections of the events, the user actions within the software application, such as entering the print routine, or the keystrokes that caused the error. This is very often difficult to accomplish since either the user may not remember all of the steps he took that caused the problem or the problem only manifests itself on the user's computer due to the configuration of the user's computer. In addition, determining the exact configuration of the user's computer is sometimes difficult since the user may not remember, for example, the type of graphics card that he installed in his computer. Therefore, it is desirable to be able to determine the configuration of a user's computer and capture information about the user's actions in order to help the Quality Assurance process.
Once a plurality of pieces of data about a machine state in a computer-based system, known as a data set, have been received from a computer-based system, it is desirable to be able to automatically process these pieces of data. In particular, it is desirable to group the data into categories of similar incidents. To categorize each piece of data, it is necessary to parse the pieces of data and automatically generate links between pieces of data that contain information about, for example, the same software crash. Duplicate pieces of data about the same event may be automatically identified and removed. It is also desirable to determine whether a particular incident is a first instance of a particular problem. A conventional bug tracking system often makes it difficult to eliminate pieces of data about the same bug or event since a person must search through all existing bug reports in order to determine the particular problem has already been located. To automatically recreate a crash, known as characterization, the system must determine the crash parameters from the piece of data and recreate the problem, if possible.
No known system automatically classifies and characterizes a data set containing information about the state of a computer-based machine. Thus, there is a need for a system and method for automatically classifying and categorizing state machine data and it is to this end that the present invention is directed.
SUMMARY OF THE INVENTION
The invention provides a system and method for automatically classifying a data set, containing pieces of data representing a state of a computer-based system, which have been generated in a variety of manners. For example, the data set may be generated by remotely monitoring a plurality of computer-based systems for the purpose of beta testing a new piece of software or by receiving a data set from a single personal computer system. In one embodiment of the invention, the information in each data set is analyzed to determine if a particular piece of data relates to a previously reported software or hardware bug, to classify the piece of data accordingly and to take some action based on the classification. For example, for duplicate pieces of data about the same bug, a counter may be incremented instead of storing the duplicate piece of data. In another example, the system may receive a data set from a variety of different computer platforms and then classify these data sets from different computer platforms.
The invention provides a system and method for automatically classifying a data set received from a computer-based system. In particular, the information in a data set which corresponds to an incident, such as a software or hardware crash, may be classified by any combination of the information in the data set. Thus, each incident may be classified based on different information in the data set. For example, the information in each piece of data is analyzed to determine if a particular piece of data relates to a previously reported software or hardware bug and classify the piece of data accordingly. If the piece of data relates to a new problem, the system may notify another system, such as a bug tracking system, or may send a message directly to the engineer responsible for the code. If the piece of data is determined to be a duplicate, duplicate pieces of data about the same bug may be eliminated, an incident counter for the particular bug may be incremented, and no additional notification is necessary.
In accordance with another aspect of the invention, the automatic classification system may perform multiple analyses for a particular data set in order to 1) automatically select the analysis which is most effective (i.e., fits most closely with the information in the data set); and/or 2) combine the multiple analyses results to generate a classification confidence level.
The invention also provides a system for automatically characterizing each bug/error identified by the automatic classification system. In particular, the system automatically attempts to reproduce an identified bug/error based on any of the information in the data set. The system may use the information in the data set to generate a template for emulating the entire environment and situation in which the incident occurred. For example, the categorization system may attempt to reproduce an error using the same size hard disk drive as the system which crashed or may use the same hardware drivers. In addition, the automatic categorization system may attempt to reproduce an error using various permutations of the crashing system's configuration to determine which factors caused the error.
In accordance with the invention, a system and method for processing data retrieved from a computer-based system is provided in which each piece of data indicates the state of each computer-based system when a predetermined event occurs within the computer-based system. In accordance with the invention, the event state data in an incident report from the computer-based s
Derossi Chris
Othmer Konstantin
Fish & Richardson P.C.
Homere Jean R.
Support.com, Inc.
LandOfFree
System and method for automatically categorizing and... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for automatically categorizing and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for automatically categorizing and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2462597