Workload aware storage platform

Patent metadata

Field	Value
Publication number	US-10579272-B2
Application number	US-201515563688-A
Country	US
Kind code	B2
Filing date	Apr 7, 2015
Priority date	Apr 7, 2015
Publication date	Mar 3, 2020
Grant date	Mar 3, 2020

Abstract

A node is provided. The node includes a receiver configured to receive data from an application. The data includes at least one of an expected workload on a dataset by the application and at least one expected characteristic of the dataset. The node includes a processor. The node includes a memory containing instructions, executable by the processor, that cause the processor to: compare the received data to a multiple storage backend configurations, select a storage backend configuration from the multiple storage backend configurations based on the comparison, cause creation of the dataset in a first storage backend having the selected storage backend configuration, determine performance data associated with at least one of performance of the first storage backend and actual workload on the dataset by the application, and modify at least one storage parameter of the first storage backend based on the determined performance data.

First claim

What is claimed is:

1. A node, comprising:
a communication interface, the communication interface configured to receive data from an application, the data including at least one of an expected workload on a dataset by the application and at least one expected characteristic of the dataset;
a processor; and
a memory, the memory containing instructions, executable by the processor, that cause the processor to:
compare the received data including at least one of the expected workload on the dataset by the application and the at least one expected characteristic of the dataset to a plurality of storage backend configurations;
select a storage backend configuration from the plurality of storage backend configurations based on the comparison;
cause creation of the dataset in an instance of a first storage backend having the selected storage backend configuration, the causing the creation including:
determining whether the instance of the first storage backend has available resources to support the dataset;
in response to determining the instance of the first storage backend lacks available resources to support the dataset, determining whether the instance of the first storage backend can be dynamically extended by increasing resources of the first storage backend; and
in response to determining the resources of the first storage backend cannot be dynamically extended, causing creation of a new software instance of the first storage backend having available resources to support the dataset, the new software instance of the first storage backend being independent of the instance of the first storage backend;

determine performance data associated with at least one of performance of the instance of the first storage backend and actual workload on the dataset by the application; and
modify at least one storage parameter of the instance of the first storage backend based on the determined performance data.

2. The node of claim 1 , wherein the at least one expected characteristic of the dataset includes at least one of storage type, expected size and requested durability.

3. The node of claim 1 , wherein each storage backend configuration defines a plurality of backend parameters.

4. The node of claim 3 , wherein the plurality of backend parameters includes at least two of a degree of randomness, read and write bias, block size, Input/Output Operations per Second (IOPS), latency, throughput, dataset size and durability.

5. The node of claim 1 , wherein causing creation of the dataset in the instance of the first storage backend having the selected storage backend configuration includes:
selecting resources for deploying the instance of the first storage backend; and
deploying the instance of the first storage backend.

6. The node of claim 1 , wherein causing creation of the dataset in the instance of the first storage backend having the selected storage backend configuration further includes:
in response to determining the resources in the instance of the first storage backend can be dynamically extended, allocating additional resources to the first storage backend to support the creation of the dataset.

7. The node of claim 1 , wherein the new software instance of the first storage backend has the selected storage backend configuration.

8. The node of claim 1 , wherein causing creation of the dataset in the instance of the first storage backend having the selected storage backend configuration further includes:
in response to determining the resources of the instance of the first storage backend cannot be dynamically extended, selecting a second storage backend configuration to support the dataset, the second storage backend configuration being less optimized to support the dataset than the selected storage backend configuration.

9. The node of claim 1 , wherein causing creation of the dataset in the instance of the first storage backend further includes:
in response to determining the instance of the first storage backend has sufficient available resources to support the dataset, configuring the available resources for the dataset.

10. The node of claim 1 , wherein the at least one storage parameter includes at least one of: number of storage tiers for the dataset, number of computing nodes, size of cache, number of interfaces and percentage of active dataset in each tier for the dataset.

11. The node of claim 1 , wherein the dataset is stored in at least one of a first tier and a second tier of the instance of the first storage backend, the first tier providing greater data access performance than the second tier;
the determined performance data indicating a percentage of active dataset of the dataset has increased; and
the modifying of the at least one parameter includes increasing the percentage of active dataset in the first tier of the instance of the first storage backend.

12. The node of claim 1 , wherein the memory further includes instructions, executable by the processor, that cause the processor to:
determine additional performance data associated with at least one of performance of the instance of the first storage backend and actual workload on the dataset by the application, the additional performance data indicating at least one expected characteristic of the dataset has changed; and
cause migration of the dataset to a second storage backend based at least in part on the determined additional performance data.

13. A method, comprising:
receiving data from an application, the data including at least one of an expected workload on a dataset by the application and at least one expected characteristic of the dataset;
comparing the received data including at least one of the expected workload on the dataset by the application and the at least one expected characteristic of the dataset to a plurality of storage backend configurations;
selecting a storage backend configuration from the plurality of storage backend configurations based on the comparison;
causing creation of the dataset in an instance of a first storage backend having the selected storage backend configuration, the causing the creation including:
determining whether the instance of the first storage backend has available resources to support the dataset;
in response to determining the instance of the first storage backend lacks available resources to support the dataset, determining whether the instance of the first storage backend can be dynamically extended by increasing resources of the first storage backend; and
in response to determining the resources of the first storage backend cannot be dynamically extended, causing creation of a new software instance of the first storage backend having available resources to support the dataset, the new software instance of the first storage backend being independent of the instance of the first storage backend;

determining performance data associated with at least one of performance of the instance of the first storage backend and actual workload on the dataset by the application; and
modifying at least one storage parameter of the instance of the first storage backend based on the determined performance data.

14. The method of claim 13 , wherein the at least one characteristic of the dataset includes at least one of storage type, expected size and requested durability.

15. The method of claim 13 , wherein each storage backend configuration defines a plurality of backend parameters.

16. The method of claim 15 , wherein the plurality of backend parameters includes at least two of a degree of randomness, read and write bias, block size, Input/Output Operati

Assignees

• Ericsson Telefon Ab L M

Inventors

• Tremblay Benoit
• Kozubal Karol
• Li Wubin

Classifications

• G06F3/0631

  by allocating resources to storage systems · CPC title

• G06F3/061

  Improving I/O performance · CPC title

• G06F3/0611 Primary

  in relation to response time · CPC title

• G06F3/0689

  Disk arrays, e.g. RAID, JBOD · CPC title

• H04L67/1097

  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] · CPC title