US 11,237,802 B1Granted

Architecture diagram analysis tool for software development

AssigneeBank of America Corporation

InventorsMadhuSudhanan Krishnamoorthy, Sreeram Raghavan

Life of this patent

Estimating the term…

Loading patent…

Abstract

A device configured to obtain an architecture diagram that includes features that are configured to form a workflow for a computer system. The device is further configured to identify the features within the architecture diagram and their metadata. The device is further configured to convert the features into vector points based on the metadata and to generate a vector map that associates vector points with their metadata. The device is further configured to input the vector points into a machine learning model and to obtain classification results for the vector points. The device is further configured to identify non-compliant features that correspond with vector points that are associated with a non-compliant classification. The device is further configured to identify alternative features for the non-compliant features, to update the vector map with the alternative features, and to update the architecture diagram based on the updated vector map.

Claims20

Dependent claims

An architecture diagram analysis device, comprising:

a memory operable to store a machine learning model configured to classify an input as either compliant or non-compliant;
and
a processor operably coupled to the memory, configured to:
obtain an architecture diagram comprising graphical representations of a plurality of features that are configured to form a workflow for a computer system, wherein the plurality of features comprises:
a plurality of components comprising software applications and hardware components;
and
communication channels between the plurality of components;
identify the plurality of features within the architecture diagram;
identify metadata that is associated with the identified features, wherein:
the metadata comprise text associated with the identified features and location information for the identified features within the architecture diagram;
and
the location information comprises information identifying where a feature is located within the architecture diagram;
convert the identified features into vector points based on the identified metadata, wherein:
a vector point is a point in an n-dimensional space;
and
the number of dimensions in the n-dimensional space is based on the number of features in the architecture diagram;
generate a vector map, wherein the vector map is configured to associate a vector point with its corresponding metadata;
input the vector points into the machine learning model;
obtain classification results for the vector points in response to inputting the vector points into the machine learning model, wherein the classification results associate each vector point with one of a compliant classification and a non-compliant classification;
identify vector points that are associated with a non-compliant classification;
identify one or more non-compliant features corresponding with the vector points that are associated with a non-compliant classification;
identify alternative features for the one or more non-compliant features, wherein each alternative feature is one of a software application and a hardware component;
update the vector map by replacing the one or more non-compliant features with the alternative features;
and
update the architecture diagram based on the updated vector map, wherein updating the architecture diagram replaces graphical representations of the one or more non-compliant features with graphical representations of the alternative features in the architecture diagram.

22 more elements

An architecture diagram analysis method for a computing system infrastructure, comprising:

obtaining an architecture diagram comprising graphical representations of a plurality of features that are configured to form a workflow for a computer system, wherein the plurality of features comprises:
a plurality of components comprising software applications and hardware components;
and
communication channels between the plurality of components;
identifying the plurality of features within the architecture diagram;
identifying metadata that is associated with the identified features, wherein:
the metadata comprise text associated with the identified features and location information for the identified features within the architecture diagram;
and
the location information comprises information identifying where a feature is located within the architecture diagram;
converting the identified features into vector points based on the identified metadata, wherein:
a vector point is a point in an n-dimensional space;
and
the number of dimensions in the n-dimensional space is based on the number of features in the architecture diagram;
generating a vector map, wherein the vector map is configured to associate a vector point with its corresponding metadata;
inputting the vector points into a machine learning model that is configured to classify an input as either compliant or non-compliant;
obtaining classification results for the vector points in response to inputting the vector points into the machine learning model, wherein the classification results associate each vector point with one of a compliant classification and a non-compliant classification;
identifying vector points that are associated with a non-compliant classification;
identifying one or more non-compliant features corresponding with the vector points that are associated with a non-compliant classification;
identifying alternative features for the one or more non-compliant features, wherein each alternative feature is one of a software application and a hardware component;
updating the vector map by replacing the one or more non-compliant features with the alternative features;
and
updating the architecture diagram based on the updated vector map, wherein updating the architecture diagram replaces graphical representations of the one or more non-compliant features with graphical representations of the alternative features in the architecture diagram.

19 more elements

A computer program comprising executable instructions stored in a non-transitory computer readable medium that when executed by a processor causes the processor to:

obtain an architecture diagram comprising graphical representations of a plurality of features that are configured to form a workflow for a computer system, wherein the plurality of features comprises:
a plurality of components comprising software applications and hardware components;
and
communication channels between the plurality of components;
identify the plurality of features within the architecture diagram;
identify metadata that is associated with the identified features, wherein:
the metadata comprise text associated with the identified features and location information for the identified features within the architecture diagram;
and
the location information comprises information identifying where a feature is located within the architecture diagram;
convert the identified features into vector points based on the identified metadata, wherein:
a vector point is a point in an n-dimensional space;
and
the number of dimensions in the n-dimensional space is based on the number of features in the architecture diagram;
generate a vector map, wherein the vector map is configured to associate a vector point with its corresponding metadata;
input the vector points into a machine learning model that is configured to classify an input as either compliant or non-compliant;
obtain classification results for the vector points in response to inputting the vector points into the machine learning model, wherein the classification results associate each vector point with one of a compliant classification and a non-compliant classification;
identify vector points that are associated with a non-compliant classification;
identify one or more non-compliant features corresponding with the vector points that are associated with a non-compliant classification;
identify alternative features for the one or more non-compliant features, wherein each alternative feature is one of a software application and a hardware component;
update the vector map by replacing the one or more non-compliant features with the alternative features;
and
update the architecture diagram based on the updated vector map, wherein updating the architecture diagram replaces graphical representations of the one or more non-compliant features with graphical representations of the alternative features in the architecture diagram.

19 more elements

Dependencies are inferred from the claim text.