Browse Prior Art Database

Storage System Level Resource Auto-rebalance

IP.com Disclosure Number: IPCOM000239626D
Publication Date: 2014-Nov-20
Document File: 7 page(s) / 137K

Publishing Venue

The IP.com Prior Art Database

Abstract

This disclosure is going to expound a method to automatically balance resource among storage pools so as to make full use of current existing resource in a storage system. The idea of “Storage System Level Resource Auto-rebalance” is to move out arrays with little IO from the source pool and then add them to a target pool with heavy IO. By this way, storage resource is better used and considerable performance increase is expectable for the target pool with limited cost of source pool performance. This function is based on some current auto-tiering technology including IO monitor and data migration.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 33% of the total text.

Page 01 of 7

Storage System Level Resource Auto

Storage System Level Resource Auto-

Introduction

Auto-tiering technology is quite popular in storage industry these days. A hybrid storage pool may contain three types of tier: High performance tier(usually built up with SSD drives), performance tier(ENT Drives) and capacity tier (NL Drives). The auto-tiering technology allows customer data to be automatically relocated to proper tier according to IO monitor result so as to get higher performance.

Almost all storage manufacturer has implemented their own function. However, all current auto-tiering technologies are restricted by storage pools(some time called raid array groups). This is to say all data migration should be within certain storage pool. When there is unbalance among storage pools, there is no mechanism to deal with such situation.

This disclosure is going to expound a method to automatically balance resource among storage pools so as to make full use of current existing resource ina storage system.

Methods

The idea of "Storage System Level Resource Auto-rebalance" is to move out arrays with little IO from the source pool and then add them to a target pool with heavy IO. By this way, storage resource is better used and considerable performance increase is expectable for the target pool with limited cost of source pool performance. This function is based on some current auto-tiering technology including IO monitor and data migration.

Cold array

Cold array

candidate

An array candidate should be an unused array or it can meet all following requirement.

1) space usage < 10% array capacity ( the percentage rate is configurable through configure file)

2) Array IO heat < 10% array IO threshold( the percentage rate is configurable through configure file)

3) It is able to move the array from the source pool. And after the array is moved out from the source pool, no array in the source pool would exceed IO threshold.

1

--rebalance

rebalance

System level Resource auto

System level Resource auto-

--rebalance

rebalance



Page 02 of 7

All arrays which meet the requirements above will be selected out and added to cold array candidate list. When there are multiple available arrays, array candidate priority will be defined. Among different types of arrays, SSD ranks has the highest priority, then ENT ranks. NL ranks' priority is the lowest. For arrays of the same type, array with lower heat has the higher priority.

Add array action will be monitored . An array newly added to a storage pool will not be chosen as a candidate array until a long period of time (defined in configure file) later.

Hot pool candidate

Two types of storage pools are qualified.

1)In this pool, there is array over IO threshold and other arrays' heat is over 50% of their threshold.( the percentage rate is configurable through configure file)

For this kind of pool. All kinds for arrays are acceptable. The add order is according to array candidate priority.


2) There is array over 80%...