Database simulation modeling framework

Patent metadata

Field	Value
Publication number	US-11907096-B2
Application number	US-202117565815-A
Country	US
Kind code	B2
Filing date	Dec 30, 2021
Priority date	Dec 30, 2021
Publication date	Feb 20, 2024
Grant date	Feb 20, 2024

Abstract

Methods, systems, and computer program products are provided for creating a resource management testing environment. An initial population of databases is established in a database ring, having an in initial count of databases and different types of databases that are determined based on an initial database population model. The initial population model receives ring classification information for the database ring from a ring grouping model. A sequence of database population-change events is generated based on a model, to change the population of the databases over time in the ring. An orchestration framework performs testing of resource manager operations based on the model-defined initial population of databases and the model-defined populations of databases changed over time. Model-defined resource usage metrics for each database are utilized to test the resource manager operations. Resource usage metrics and database add/drop events of a production system are used to train the models.

First claim

What is claimed is:

1. A system for creating a resource management testing environment, the system comprising:
one or more processors; and
one or more memory devices that store program code to be executed by the one or more processors, the program code comprising:
an initial database (DB) population generator configured to:
establish a model-defined initial population of databases in a database ring of a database system, the model-defined initial population of databases comprising a model-defined count of databases and proportions of different model-defined types of databases based on output of an initial database population model;

a DB population-change generator configured to:
generate a model-defined sequence of database population-change events that are used to change the population of the databases in the database ring over time based on output of a database population-change events model, wherein the database population-change events model is trained based on production telemetry data to mimic behavior of a production system; and

an orchestration framework configured to:
perform a test of resource management in the database system based on the model-defined initial population of databases and the model defined populations of databases changed over time based on the output of the database population-change events model.

2. The system of claim 1 , wherein at least one of the following occurs according to the model-defined sequence of database population-change events:
the model-defined count of databases within the database ring changes over time; or
the proportions of the different model-defined types of databases within the database ring changes over time.

3. The system of claim 1 , wherein a type of database of the different model-defined types of databases is determined based on one or more of the following:
a type of software utilized for the database,
a type of hardware utilized for the database,
a generation of software utilized for the database,
a generation of hardware utilized for the database,
a configuration of software components of the database,
a configuration of hardware components of the database,
available disk capacity,
available memory capacity,
a number of compute cores, and
remote or local database storage type.

4. The system of claim 1 , wherein the model-defined sequence of database population-change events comprises at least one of a create database event or a drop database event.

5. The system of claim 1 , wherein the program code further comprises a database ring grouping generator configured to:
generate a plurality of model-defined ring groups comprising at least the database ring based on output of a ring grouping model trained based on resource utilization metrics of a production database system,
wherein the database ring belongs to a classification that is based on at least one resource utilization metric dimension and at least one resource utilization metric level type defined based on the output of the ring grouping model.

6. The system of claim 1 , wherein the program code further comprises a database resource usage generator configured to:
generate a model-defined set of database resource usage metrics for a database in a current population of databases; wherein:
the current population of databases comprises one of the initial population of databases or a subsequent population of databases, and
the database in the current population of databases reports the model-defined set of database resource usage metrics to the orchestration framework for testing at least one of database placement, load balancing, or scaling operations in the database ring by the orchestration framework in response to the reported model-defined set of database resource usage metrics.

7. The system of claim 6 , wherein the model-defined set of database resource usage metrics comprises at least one of model-defined compute resource usage, model-defined memory usage, or model-defined disk usage and models utilized for defining the set of database resource usage metrics are trained based on resource usage metrics of a production database system.

8. A system, for creating a resource management testing environment, the system comprising:
one or more processors; and
one or more memory devices that store program code to be executed by the one or more processors, the program code comprising:
an initial database (DB) population generator configured to:
establish a model-defined initial population of databases in a database ring where each database in the initial population has an artificial initial resource utilization metric value determined based on output of an initial database population model, wherein the initial database population model is trained based on resource usage metrics of a production database system;
wherein the initial population of databases in the database ring comprises a count of databases having different types of databases that are determined based on the initial database population model; and

an orchestration framework configured to:
perform a test of a resource manager for the database ring utilizing the artificial initial resource utilization metric value, wherein the resource manager receives the artificial initial resource utilization metric value from at least one database of the model defined initial population of databases.

9. The system of claim 8 , wherein types of the different types of databases in the initial population comprise local-storage databases or remote-storage databases determined based on the initial database population model.

10. The system of claim 8 , wherein types of the different types of databases in the initial population comprise different resource capacities or different resource counts in the initial population of databases determined based on the initial database population model.

11. The system of claim 8 , wherein the initial database population model receives input comprising one or more test scenario parameters for:
target resource utilization levels,
target range of number of databases, and
target ring group classification,

for defining the initial population of databases and each artificial initial resource utilization metric value for each database of the initial population of databases.

12. The system of claim 8 , wherein the database ring is classified according to at least one type of resource usage metric level determined based on output of a ring grouping model, wherein the ring grouping model is trained based on database resource usage metrics of a production database system and outputs a definition for at least one ring group comprising the database ring.

13. The system of claim 8 , wherein the orchestration framework is further configured to:
perform a test of the resource manager for the database ring utilizing a subsequent artificial resource utilization metric value of at least one database of a subsequent model defined population of databases; and
perform a test of the resource manager in the database system based on a sequence of model defined database population-change events.

14. A system for managing a resource management testing environment, the system comprising:
one or more processors; and
one or more memory devices that store program code to be executed by the one or more processors, the program code comprising:
a database (DB) population-change generator configured to:
generate a model-defined sequence of database population-change events that change a population of databases in a database ring over time based on output of a database population-change events model, wherein the dat

Assignees

• Microsoft Technology Licensing Llc

Inventors

• Ye Zi
• Moeller Justin Grant
• Lin Ya
• Lang Willis

Classifications

• G06F11/3698

  Environments for analysis, debugging or testing of software · CPC title

• G06F11/3433 Primary

  for load management (allocation of a server based on load conditions G06F9/505; load rebalancing G06F9/5083; redistributing the load in a network by a load balancer H04L67/1029) · CPC title

• G06F11/3006

  where the computing system is distributed, e.g. networked systems, clusters, multiprocessor systems (multiprogramming arrangements G06F9/46; allocation of resources G06F9/50) · CPC title

• G06F11/3457

  Performance evaluation by simulation · CPC title

• G06F16/217

  Database tuning (G06F16/2282 takes precedence; database performance monitoring G06F11/3409) · CPC title