# US 2021/0240517 A1 · Data analysis for predictive scaling of container(s) based on prior user transaction(s)

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

- Number: US 2021/0240517 A1
- Document ID: US20210240517A1
- Status: Published application
- Kind: A1
- Application number: 16/782,434
- Filed: 2020-02-05
- Published: 2021-08-05
- Earliest claimed: 2020-02-05
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Biswajit MOHAPATRA; Prasanta Kumar PAL; Venkata Vinay Kumar PARISA
- Web page: https://patentgrep.com/patent/US20210240517A1

## CPC

- G06F 9/45558: {Hypervisor-specific management and integration aspects}
- G06F 9/466: {Transaction processing}
- G06F 9/445: Program loading or initiating (bootstrapping G06F9/4401; security arrangements for program loading or initiating G06F21/57)
- G06F 9/5027: {the resource being a machine, e.g. CPUs, Servers, Terminals}
- G06N 3/09: Supervised learning
- G06N 20/00: Machine learning
- G06F 9/5005: {to service a request}
- G06F 9/505: {considering the load}
- G06F 2009/4557: {Distribution of virtual machine instances; Migration and load balancing}
- G06F 2009/45575: {Starting, stopping, suspending or resuming virtual machine instances}
- G06F 2209/5014: Reservation
- G06F 2209/5019: Workload prediction
- G06F 2209/508: Monitor
- G06N 3/006: based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
- G06N 3/08: Learning methods
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

Predictive scaling of containers is provided based on obtaining, by one or more processors, prior transaction data on one or more of a user's prior transactions on one or more transaction-centric applications, and analyzing by the processor(s), the prior transaction data to predict whether a current transaction of the user on an application is deterministic or non-deterministic as to one or more functional layers of multiple functional layers of the application. Based on predicting that the current transaction is deterministic as to the functional layer(s) of the application, the processor(s) determines compute resource requirements for the deterministic transaction, and launches one or more containers in advance to run the functional layer(s) based, at least in part, on the computed resource requirements of the deterministic transaction.

## Claims

### Claim 1

A computer-implemented method comprising:
obtaining, by one or more processors, prior transaction data on one or more of a user's prior transactions on one or more transaction-centric applications;
analyzing, by the one or more processors, the prior transaction data to predict whether a current transaction of the user on an application is deterministic or non-deterministic as to one or more functional layers of multiple functional layers of the application; and
based on predicting that the user transaction is deterministic as to the one or more functional layers of the application:
determining, by the one or more processors, compute resource requirements for the deterministic transaction; and
launching, by the one or more processors, one or more containers in advance to run the one or more functional layers based, at least in part, on the computed resource requirements of the deterministic transaction.

### Claim 2 (depends on 1)

The computer-implemented method of claim 1, further comprising, based on predicting that the current transaction is non-deterministic as to the one or more functional layers of the application, holding off determining compute resource requirements for the non-deterministic transaction, and holding off launching one or more containers in advance to run the one or more functional layers.

### Claim 3 (depends on 2)

The computer-implemented method of claim 2, further comprising:
based on predicting, by the one or more processors, that the current transaction is non-deterministic, obtaining, by the one or more processors, transaction data on the current transaction on the application, and using, by the one or more processors, the transaction data and the prior transaction data to subsequently predict that the non-deterministic transaction has become deterministic as to at least one functional layer of the application, and based on the current transaction becoming deterministic as to the at least one functional layer of the application, launching, by the one or more processors, at least one container to ensure availability of the at least one functional layer of the application at time of demand.

### Claim 4 (depends on 1)

The computer-implemented method of claim 1, further comprising, based on predicting that the current transaction is deterministic, dynamically providing, by the one or more processors, one or more auto-scaling rules to guide the launching of the one or more containers to ensure availability at time of demand of the one or more functional layers of the application.

### Claim 5 (depends on 1)

The computer-implemented method of claim 1, wherein the analyzing, by the one or more processors, comprises using a probabilistic model to predict whether the current transaction is likely to use the one or more functional layers of the application, and thereby predict whether the current transaction is deterministic or non-deterministic as to the one or more functional layers of the application.

### Claim 6 (depends on 1)

The computer-implemented method of claim 1, where the prior transaction data comprises trend data representative of the user's N-most recent transactions on the one or more transaction-centric applications, and the analyzing comprises using, by the one or more processors, the trend data and a probabilistic model to predict whether the current transaction is deterministic or non-deterministic as to one or more functional layers of the application.

### Claim 7 (depends on 6)

The computer-implemented method of claim 6, wherein the trend data comprises prior transaction data on functional layers accessed during the N-most recent user transactions on the one or more transaction-centric applications, where N≥2.

### Claim 8 (depends on 1)

The computer-implemented method of claim 1, further comprising detecting, by the one or more processors, initiating of the current transaction on the application, and performing the obtaining and the analyzing based on detecting the initiating of the current transaction on the application.

### Claim 9 (depends on 8)

The computer-implemented method of claim 8, wherein the detecting comprises detecting access by the current transaction of an initial functional layer of the multiple functional layers of the application, and wherein the one or more functional layers are one or more subsequent functional layers of the multiple functional layers of the application, each functional layer of the multiple functional layers running in one or more respective containers.

### Claim 10 (depends on 9)

The computer-implemented method of claim 9, wherein a functional layer of the multiple functional layers comprises a respective microservice running in a respective container.

### Claim 11 (depends on 1)

The computer-implemented method of claim 1, further comprising repeating the obtaining and the analyzing across a current plurality of user transactions on the application, the repeating correlating current users' prior transaction data on the one or more transaction-centric applications to a predictive launching of containers in advance for running the one or more functional layers of the application based on the current plurality of user transactions.

### Claim 12

A system comprising:
a memory;
one or more processors in communication with the memory; and
program instructions executable by the one or more processors via the memory to perform a method, the method comprising:
obtaining, by one or more processors, prior transaction data on one or more of a user's prior transactions on one or more transaction-centric applications;
analyzing, by the one or more processors, the prior transaction data to predict whether a current transaction of the user on an application is deterministic or non-deterministic as to one or more functional layers of multiple functional layers of the application; and
based on predicting that the user transaction is deterministic as to the one or more functional layers of the application:
determining, by the one or more processors, compute resource requirements for the deterministic transaction; and
launching, by the one or more processors, one or more containers in advance to run the one or more functional layers based, at least in part, on the computed resource requirements of the deterministic transaction.

### Claim 13 (depends on 12)

The system of claim 12, further comprising, based on predicting that the current transaction is non-deterministic as to the one or more functional layers of the application, holding off determining compute resource requirements for the non-deterministic transaction, and holding off launching one or more containers in advance to run the one or more functional layers.

### Claim 14 (depends on 12)

The system of claim 12, wherein the analyzing, by the one or more processors, comprises using a probabilistic model to predict whether the current transaction is likely to use the one or more functional layers of the application, and thereby predict whether the current transaction is deterministic or non-deterministic as to the one or more functional layers of the application.

### Claim 15 (depends on 12)

The system of claim 12, where the prior transaction data comprises trend data representative of the user's N-most recent transactions on the one or more transaction-centric applications, and the analyzing comprises using, by the one or more processors, the trend data and a probabilistic model to predict whether the current transaction is deterministic or non-deterministic as to one or more functional layers of the application.

### Claim 16 (depends on 15)

The system of claim 15, wherein the trend data comprises prior transaction data on functional layers accessed during the N-most recent user transactions on the one or more transaction-centric applications, where N≥2.

### Claim 17 (depends on 12)

The system of claim 12, further comprising repeating the obtaining and the analyzing across a current plurality of user transactions on the application, the repeating correlating current users' prior transaction data on the one or more transaction-centric applications to a predictive launching of containers in advance for running the one or more functional layers of the application based on the current plurality of user transactions.

### Claim 18

A computer program product comprising:
a computer-readable storage medium having program instructions embodied therewith, the program instructions being executable by one or more processors to cause the one or more processors to:
obtain, by the one or more processors, prior transaction data on one or more of a user's prior transactions on one or more transaction-centric applications;
analyze, by the one or more processors, the prior transaction data to predict whether a current transaction of the user on an application is deterministic or non-deterministic as to one or more functional layers of multiple functional layers of the application; and
based on predicting that the user transaction is deterministic as to the one or more functional layers of the application:
determine, by the one or more processors, compute resource requirements for the deterministic transaction; and
launch, by the one or more processors, one or more containers in advance to run the one or more functional layers based, at least in part, on the computed resource requirements of the deterministic transaction.

### Claim 19 (depends on 18)

The computer program product of claim 18, wherein the program instructions are further executable by the one or more processors to cause the one or more processors to hold off, based on predicting that the current transaction is non-deterministic as to the one or more functional layers of the application, determining compute resource requirements for the non-deterministic transaction, and to hold of launching of the one or more containers in advance to run the one or more functional layers.

### Claim 20 (depends on 18)

The computer program product of claim 18, wherein the program instructions are further executable by the one or more processors to cause the one or more processors to use a probabilistic model in analyzing the prior transaction data to predict whether the current transaction is likely to use the one or more functional layers of the application, and thereby to predict whether the current transaction is deterministic or non-deterministic as to the one or more functional layers of the application.
