# US 2021/0342310 A1 · Cognitive Method to Perceive Storages for Hybrid Cloud Management

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

- Number: US 2021/0342310 A1
- Document ID: US20210342310A1
- Status: Published application
- Kind: A1
- Application number: 16/865,395
- Filed: 2020-05-03
- Published: 2021-11-04
- Earliest claimed: 2020-05-03
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Jun Yang; Hui Lin; Peng Fei Tian; Yue Wang
- Web page: https://patentgrep.com/patent/US20210342310A1

## CPC

- G06F 11/3006: {where the computing system is distributed, e.g. networked systems, clusters, multiprocessor systems (multiprogramming arrangements G06F9/46; allocation of resources G06F9/50)}
- G06F 16/1827: {Management specifically adapted to NAS (management of storage area networks [SAN] G06F3/067)}
- G06F 3/0604: {Improving or facilitating administration, e.g. storage management}
- G06F 3/061: {Improving I/O performance}
- G06F 3/0653: {Monitoring storage devices or systems}
- G06F 3/067: {Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]}
- G06F 3/0685: {Hybrid storage combining heterogeneous device types, e.g. hierarchical storage, hybrid arrays}
- G06F 7/5443: {Sum of products (for applications thereof, see the relevant places, e.g. G06F17/10, H03H17/00)}
- G06F 11/3034: {where the computing system component is a storage system, e.g. DASD based or network based (digital input from or digital output to record carriers G06F3/06; digital recording or reproducing G11B20/18; for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS], H04L67/1097)}
- G06F 11/3051: {Monitoring arrangements for monitoring the configuration of the computing system or of the computing system component, e.g. monitoring the presence of processing resources, peripherals, I/O links, software programs (verification or detection of system hardware configuration G06F11/2247)}
- G06F 16/1734: {Details of monitoring file system events, e.g. by the use of hooks, filter drivers, logs}
- G06N 3/0499: Feedforward networks
- G06N 3/08: Learning methods
- G06N 3/09: Supervised learning
- G06Q 10/06: Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q 30/0283: Price estimation or determination
- G06N 3/048: Activation functions
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR

## Abstract

A mechanism is provided in a data processing system for hybrid cloud management. The mechanism generates hybrid cloud storage features and hybrid cloud environment factors. The mechanism performs a dynamic confidence method on the hybrid cloud features based on the hybrid cloud environment factors using a deep learning model to generate a hybrid cloud storage profile. The mechanism performing model optimization on the deep learning model and generating a files-storage matrix. The mechanism generates a hybrid cloud file profile based on the hybrid cloud storage profile and the files-storage matrix. The mechanism generates a target file matrix based on the hybrid cloud storage profile and the hybrid cloud file profile. The mechanism stores files based on the target file matrix.

## Claims

### Claim 1

A method, in a data processing system, for hybrid cloud management, the method comprising:
generating hybrid cloud storage features and hybrid cloud environment factors;
performing a dynamic confidence method on the hybrid cloud features based on the hybrid cloud environment factors using a deep learning model to generate a hybrid cloud storage profile, wherein the hybrid cloud storage profile provides, for each storage, a prediction for each combination of category and genre;
performing model optimization on the deep learning model;
generating a files-storage matrix;
generating a hybrid cloud file profile based on the hybrid cloud storage profile and the files-storage matrix, wherein the hybrid cloud profile provides, for each file, a prediction for each combination of category and genre;
generating a target file matrix based on the hybrid cloud storage profile and the hybrid cloud file profile, wherein the target file matrix provides, for each file, a prediction for each storage; and
storing files based on the target file matrix.

### Claim 2 (depends on 1)

The method of claim 1, wherein the deep learning model classifies the hybrid cloud features and hybrid cloud environment factors into a category and genre.

### Claim 3 (depends on 2)

The method of claim 2, wherein the category is public cloud storage, private cloud storage, or premise storage.

### Claim 4 (depends on 2)

The method of claim 2, wherein the genre is memory, solid-state device (SSD) storage, local disks, software-defined network (SDN) storage, or public cloud storage.

### Claim 5 (depends on 1)

The method of claim 1, wherein the hybrid cloud storage features comprise space, timestamp of space, data center, site, genre, read/write, cost, security, special policy.

### Claim 6 (depends on 1)

The method of claim 1, wherein performing the dynamic confidence method comprises:
concatenating the hybrid cloud environment factors and the hybrid cloud features to generate concatenated features;
performing an element-wise add on the concatenated features and storage features;
performing a sigmoid function on a result of the element-wise add; and
performing an element-wise product on a result of the sigmoid function and the storage features.

### Claim 7 (depends on 1)

The method of claim 1, wherein performing model optimization comprises:
calculating an initial weight of each predictor is as follows:
w
i
=
1
m
,
where m is the number of instances;
calculating a weighted error as follows:
r
j
=
∑
i
=
1
(
y
^
i
≠
y
i
)
m
⁢
w
i
∑
i
=
1
m
⁢
w
i
,
where ŷi is a predicted result and yi is an actual value;
calculating a weight of each model as follows:
α
j
=
β
*
log
⁢
1
-
r
j
r
j
,
where β is a learning rate of the model;
updating the weight as follows:
for
⁢
⁢
i
=
1
,
2
,
…
⁢
,
m
,
w
i
=
{
w
i
if
⁢
⁢
y
^
i
=
y
i
w
i
⁢
exp
⁡
(
α
j
)
if
⁢
⁢
y
^
i
≠
y
i
generating a final prediction result based on a joint prediction result of all models as follows:
ŷi(x)=max(Σj=1Nαj) for ŷj(x)=k,
where N is a total number of models and k is the predicted value.

### Claim 8 (depends on 1)

The method of claim 1, wherein generating the hybrid cloud file profile comprises performing a dot product on the hybrid cloud storage profile and the hybrid cloud file matrix.

### Claim 9 (depends on 1)

The method of claim 1, wherein generating the target file matrix comprises performing a cosine similarity on the hybrid cloud storage profile and the hybrid cloud file profile.

### Claim 10 (depends on 1)

The method of claim 1, wherein generating the files-storage matrix comprises:
performing analytics to obtain file-storage statistics; and
generating the files-storage matrix based on the file-storage statistics.

### Claim 11

A computer program product comprising a computer readable storage medium having a computer readable program stored therein, wherein the computer readable program, when executed on a computing device, causes the computing device to:
generate hybrid cloud storage features and hybrid cloud environment factors;
perform a dynamic confidence method on the hybrid cloud features based on the hybrid cloud environment factors using a deep learning model to generate a hybrid cloud storage profile, wherein the hybrid cloud storage profile provides, for each storage, a prediction for each combination of category and genre;
perform model optimization on the deep learning model;
generate a files-storage matrix;
generate a hybrid cloud file profile based on the hybrid cloud storage profile and the files-storage matrix, wherein the hybrid cloud profile provides, for each file, a prediction for each combination of category and genre;
generate a target file matrix based on the hybrid cloud storage profile and the hybrid cloud file profile, wherein the target file matrix provides, for each file, a prediction for each storage; and
store files based on the target file matrix.

### Claim 12 (depends on 11)

The computer program product of claim 11, wherein the deep learning model classifies the hybrid cloud features and hybrid cloud environment factors into a category and genre.

### Claim 13 (depends on 12)

The computer program product of claim 12, wherein the category is public cloud storage, private cloud storage, or premise storage.

### Claim 14 (depends on 12)

The computer program product of claim 12, wherein the genre is memory, solid-state device (SSD) storage, local disks, software-defined network (SDN) storage, or public cloud storage.

### Claim 15 (depends on 11)

The computer program product of claim 11, wherein the hybrid cloud storage features comprise space, timestamp of space, data center, site, genre, read/write, cost, security, special policy.

### Claim 16 (depends on 11)

The computer program product of claim 11, wherein performing the dynamic confidence method comprises:
concatenating the hybrid cloud environment factors and the hybrid cloud features to generate concatenated features;
performing an element-wise add on the concatenated features and storage features;
performing a sigmoid function on a result of the element-wise add; and
performing an element-wise product on a result of the sigmoid function and the storage features.

### Claim 17 (depends on 11)

The computer program product of claim 11, wherein performing model optimization comprises:
calculating an initial weight of each predictor is as follows:
w
i
=
1
m
,
where m is the number of instances;
calculating a weighted error as follows:
r
j
=
∑
i
=
1
(
y
^
i
≠
y
i
)
m
⁢
w
i
∑
i
=
1
m
⁢
w
i
,
where ŷi is a predicted result and yi is an actual value;
calculating a weight of each model as follows:
α
j
=
β
*
log
⁢
1
-
r
j
r
j
,
where β is a learning rate of the model;
updating the weight as follows:
for
⁢
⁢
i
=
1
,
2
,
…
⁢
,
m
,
w
i
=
{
w
i
if
⁢
⁢
y
^
i
=
y
i
w
i
⁢
exp
⁡
(
α
j
)
if
⁢
⁢
y
^
i
≠
y
i
generating a final prediction result based on a joint prediction result of all models as follows:
ŷi(x)=max(Σj=1Nαj) for ŷj(x)=k,
where N is a total number of models and k is the predicted value.

### Claim 18 (depends on 11)

The computer program product of claim 11, wherein generating the hybrid cloud file profile comprises performing a dot product on the hybrid cloud storage profile and the hybrid cloud file matrix.

### Claim 19 (depends on 11)

The computer program product of claim 11, wherein generating the target file matrix comprises performing a cosine similarity on the hybrid cloud storage profile and the hybrid cloud file profile.

### Claim 20

An apparatus comprising:
a processor; and
a memory coupled to the processor, wherein the memory comprises instructions which, when executed by the processor, cause the processor to:
generate hybrid cloud storage features and hybrid cloud environment factors;
perform a dynamic confidence method on the hybrid cloud features based on the hybrid cloud environment factors using a deep learning model to generate a hybrid cloud storage profile, wherein the hybrid cloud storage profile provides, for each storage, a prediction for each combination of category and genre;
perform model optimization on the deep learning model;
generate a files-storage matrix;
generate a hybrid cloud file profile based on the hybrid cloud storage profile and the files-storage matrix, wherein the hybrid cloud profile provides, for each file, a prediction for each combination of category and genre;
generate a target file matrix based on the hybrid cloud storage profile and the hybrid cloud file profile, wherein the target file matrix provides, for each file, a prediction for each storage; and
store files based on the target file matrix.
