# US 2026/0086792 A1 · Optimizing software upgrade processes in data center architectures

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0086792 A1
- Document ID: US20260086792A1
- Status: Published application
- Kind: A1
- Application number: 18/976,824
- Filed: 2024-12-11
- Published: 2026-03-26
- Earliest claimed: 2024-09-25
- Assignees: Hewlett Packard Enterprise Development LP
- Applicants: Hewlett Packard Enterprise Development LP
- Inventors: Srinivas Subbarao Mulgund; Jayadev Mandya Sudarshan Kumar Jain; Rohit Bangalore Vijaya; Sravanthi Rapa
- Web page: https://patentgrep.com/patent/US20260086792A1

## CPC

- G06F 8/65: Updates (security arrangements therefor G06F21/57)
- G06F 8/71: Version control (security arrangements therefor G06F21/57); Configuration management
- G06F 11/1433: {during software upgrading}
- G06F 11/3698: Environments for analysis, debugging or testing of software
- G06N 3/044: Recurrent networks, e.g. Hopfield networks
- G06N 3/045: Combinations of networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/048: Activation functions
- G06N 3/08: Learning methods
- G06N 3/082: modifying the architecture, e.g. adding, deleting or silencing nodes or connections
- G06N 3/084: Backpropagation, e.g. using gradient descent
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- G06N 20/20: Ensemble learning
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

In certain implementations, a system includes one or more processors, and one or more non-transitory computer-readable storage media storing programming for execution by the one or more processors, the programming including instructions to obtain version information for a plurality of software and firmware sub-components of a service pack installed on a computer system, input the obtained version information into a trained neural network model, receive, from the trained neural network model, a prediction of a service pack version corresponding to the plurality of software and firmware sub-components, determine, based on the predicted service pack version, a compatible target service pack for the computer system, and initiate an upgrade process to install the compatible target service pack on the computer system.

## Claims

### Claim 1

A system, comprising:
one or more processors; and
one or more non-transitory computer-readable storage media storing programming for execution by the one or more processors, the programming comprising instructions to:
obtain version information for a plurality of software and firmware sub-components of a service pack installed on a computer system;
input the obtained version information into a trained neural network model;
receive, from the trained neural network model, a prediction of a service pack version corresponding to the plurality of software and firmware sub-components;
determine, based on the predicted service pack version, a compatible target service pack for the computer system; and
initiate an upgrade process to install the compatible target service pack on the computer system.

### Claim 2 (depends on 1)

The system of claim 1, wherein the trained neural network model is a Convolutional Neural Network (CNN) model.

### Claim 3 (depends on 2)

The system of claim 2, wherein the CNN model comprises three parallel networks, each of the three parallel networks having a different filter size from the other parallel networks.

### Claim 4 (depends on 1)

The system of claim 1, wherein the programming comprises further instructions to:
receive, from the trained neural network model, a probability score that indicates a confidence of the trained neural network model in the prediction of the service pack version.

### Claim 5 (depends on 1)

The system of claim 1, wherein inputting the obtained version information into the trained neural network model comprises:
performing a data preparation process to convert the obtained version information into a delineated string; and
converting the delineated string into tokens using a predefined data map.

### Claim 6 (depends on 1)

The system of claim 1, wherein obtaining the version information for the plurality of software and firmware sub-components of the service pack installed on the computer system comprises sending a request, via an Application Programming Interface (API), to the computer system to retrieve the version information for the plurality of software and firmware sub-components.

### Claim 7

A computer-implemented method, comprising:
obtaining version information for a plurality of software and firmware sub-components of a service pack installed on a computer system;
inputting the obtained version information into a trained neural network model;
receiving, from the trained neural network model, a prediction of a service pack version corresponding to the plurality of software and firmware sub-components;
determining, based on the predicted service pack version, a compatible target service pack for the computer system; and
initiating an upgrade process to install the compatible target service pack on the computer system.

### Claim 8 (depends on 7)

The computer-implemented method of claim 7, wherein obtaining the version information for the plurality of software and firmware sub-components of the service pack comprises:
sending a request, via an Application Programming Interface (API), to the computer system to retrieve the version information for the plurality of software and firmware sub-components.

### Claim 9 (depends on 8)

The computer-implemented method of claim 8, wherein the plurality of software and firmware sub-components of the service pack comprise at least 20 sub-components.

### Claim 10 (depends on 7)

The computer-implemented method of claim 7, wherein inputting the obtained version information into the trained neural network model comprises:
performing a first data preparation process on the obtained version information to remove invalid entries in the obtained version information;
after performing the first data preparation process, performing a second data preparation process to convert the obtained version information into a delineated string; and
converting the delineated string into tokens using a predefined data map.

### Claim 11 (depends on 10)

The computer-implemented method of claim 10, wherein inputting the obtained version information into the trained neural network model further comprises inputting the tokens into the trained neural network model.

### Claim 12 (depends on 11)

The computer-implemented method of claim 11, wherein the trained neural network model is a Convolutional Neural Network (CNN) model.

### Claim 13 (depends on 12)

The computer-implemented method of claim 12, wherein the CNN model comprises three parallel networks, each of the three parallel networks having a different filter size from the other parallel networks.

### Claim 14 (depends on 7)

The computer-implemented method of claim 7, further comprising:
determining, based on the predicted service pack version, an appropriate upgrade path to reach the compatible target service pack for the computer system.

### Claim 15

A computer-implemented method, comprising:
fetching a list of discrete sub-component versions from a server node using an Application Programming Interface (API), the server node being part of a disaggregated hyper-converged infrastructure (dHCI) system;
inputting the fetched sub-component versions into a trained three-channel convolutional neural network (CNN) model;
predicting a base service pack version installed on the server node using the trained three-channel CNN model;
determining a target service pack version and a supported upgrade path to reach the target service pack version based on the predicted base service pack version; and
updating the server node by installing the target service pack version on the server node.

### Claim 16 (depends on 15)

The computer-implemented method of claim 15, further comprising training the three-channel CNN model, the training comprising randomizing sub-component versions across supported base service pack versions to generate an augmented dataset.

### Claim 17 (depends on 16)

The computer-implemented method of claim 16, wherein training the three-channel CNN model further comprises converting the augmented dataset into tokens using a predefined data map.

### Claim 18 (depends on 15)

The computer-implemented method of claim 15, wherein each channel of the three-channel CNN model has a different filter size.

### Claim 19 (depends on 15)

The computer-implemented method of claim 15, wherein predicting the base service pack version installed on the server node using the trained three-channel CNN model comprises:
concatenating outputs from all the channels of the trained three-channel CNN model into a single concatenated output; and
passing the single concatenated output through a dense layer for classification.

### Claim 20 (depends on 15)

The computer-implemented method of claim 15, further comprising:
generating, using the trained three-channel CNN model, a probability score that indicates a confidence level of the prediction of the base service pack version; and
determining whether the probability score meets or exceeds a predetermined threshold value before proceeding with determining the target service pack version and the supported upgrade path.
