Systematic failure remediation

Abstract

Aspects of the present invention provide a tool for analyzing and remediating an update-related failure. In an embodiment, a failure state of a computer system that has been arrived at as a result of an update is captured. A semantic diff that includes the difference between the failure state and at least one of an original state or a completion state is then computed. This semantic diff is transformed into a feature vector format. Then the transformed semantic diff is analyzed to determine a remediation for the update. Failure and/or resolution signatures can be constructed using the semantic diff and contextual data, and these signatures can be used in comparison and analysis of failures and resolutions.

What is claimed is: 
     
         1 . A method for remediating an update-related failure, comprising:
 capturing a failure state of a computer system arrived at as a result of the update-related failure;   computing a semantic diff, the semantic diff being a difference between the failure state and at least one of an original state or a completion state;   transforming the semantic diff into a feature vector format; and   analyzing the transformed semantic diff to determine a remediation for the update-related failure;   constructing a signature for at least one of failures or resolutions, where a signature comprises a semantic diff and a context.   
     
     
         2 . The method of  claim 1 , wherein the failure state includes contextual data that includes at least one of: the captured state of the system before the update was performed or a description of an operation that preceded the update-related failure or an annotation describing a resolution of the update-related failure. 
     
     
         3 . The method of  claim 2 , wherein the operation includes at least one of: a modification to an operating system of the computer system, modification to an application of the computer system, modification of a management agent in the computer system, modification of settings of any software or hardware component of the computer system, or a modification to communications parameters for the computer system. 
     
     
         4 . The method of  claim 1 , wherein the feature vector format includes a vector space model. 
     
     
         5 . The method of  claim 4 , the analyzing further comprising:
 accessing a stored vector space model that represents a previously computed semantic diff; and   calculating a similarity between the vector space model and the stored vector space model using the following:   
       
         
           
             
               
                 
                   sim 
                    
                   
                     ( 
                     
                       
                         D 
                         i 
                       
                       , 
                       
                         D 
                         j 
                       
                     
                     ) 
                   
                 
                 = 
                 
                   
                     
                       ∑ 
                       k 
                     
                      
                     
                       
                         D 
                         ik 
                       
                        
                       
                         D 
                         jk 
                       
                     
                   
                   
                     
                       
                         
                           ∑ 
                           k 
                         
                          
                         
                           D 
                           ik 
                           2 
                         
                       
                     
                      
                     
                       
                         
                           ∑ 
                           k 
                         
                          
                         
                           D 
                           jk 
                           2 
                         
                       
                     
                   
                 
               
               , 
             
           
         
         where:
 k is an attribute location within the vector space model, 
 D i , is the vector space model, and 
 D j  is the stored vector space model. 
 
       
     
     
         6 . The method of  claim 1 , the analyzing further comprising:
 computing the similarity between a given semantic diff and a plurality of previously computed structural diffs;   associating the semantic diff with a most proximate previously computed semantic diff;   repeatedly associating the semantic diff with a each of a series of next proximate previously semantic diffs; and   representing the associations with respect to the semantic diff in tree form.   
     
     
         7 . The method of  claim 1 , the analyzing further comprising using a generative model on the semantic diff to derive essential features of the semantic diff that led to the update-related failure. 
     
     
         8 . The method of  claim 1 , the analyzing further comprising:
 comparing the semantic diff with a plurality of previously computed semantic diffs; and   determining at least one matching semantic diff from the previously computed semantic diffs based on the comparison;   retrieving a resolution annotation from the at least one matching semantic diff; and   resolving the migration-related failure based on the resolution annotation.   
     
     
         9 . The method of  claim 1 , wherein the update-based failure occurs as a result of a migration from a server to a cloud-based virtual server. 
     
     
         10 . The method of  claim 1 , further comprising:
 constructing the signature by comparing at least one of failures or resolutions; and   analyzing the signature by comparing the signature with a set of failures and resolutions;   wherein the signature includes both the semantic diff and a context that includes at least one of: a captured state, or annotations, or operations performed.   
     
     
         11 . A method for deploying an application for remediating an update-related failure, comprising:
 providing a computer infrastructure being operable to:
 capture a failure state of a computer system arrived at as a result of the update-related failure; 
 compute a semantic diff, the semantic diff being a difference between the failure state and at least one of an original state or a completion state; 
 transform the semantic diff into a feature vector format; and 
 analyze the transformed semantic diff to determine a remediation for the update-related failure; 
 construct a signature for at least one of failure or remediation that is used for comparison and matching.