Allocating RAID Storage Volumes Across a Distributed Network of Storage Elements

Patent metadata

Field	Value
Publication number	US-2017199695-A1
Application number	US-201715468516-A
Country	US
Kind code	A1
Filing date	Mar 24, 2017
Priority date	Mar 31, 2014
Publication date	Jul 13, 2017
Grant date	—

Abstract

A distributed network of storage elements (DNSE) is provided in which the physical capacity of each drive is split into a set of equal sized logical splits which are individually protected within the DNSE using separate RAID groups. To reduce restoration latency, members of the RAID groups having a member in common on a given drive are spread within the DNSE to minimize the number of sets of drives within the DNSE that have RAID members in common. By causing the splits to be protected by RAID groups, restoration of the splits may occur in parallel involving multiple drives within the DNSE. By minimizing the overlap between RAID members on various drives, failure of a given drive will not require multiple reads from another drive in the DNSE. Likewise, spare splits are distributed to enable write recovery to be performed in parallel on multiple drives within the DNSE.

First claim

What is claimed is:

1 - 40 . (canceled)

41 . A method of allocating storage volumes across a distributed network of storage elements, the method comprising the steps of:
defining at least two logical splits on each of the storage elements; defining protection groups for the splits of a first of the storage elements using at least two separate protection groups, each of the at least two separate protection groups encompassing a plurality of other protection group members stored on splits of other storage elements of the distributed network of storage elements; and distributing the other protection group members of the at least two separate protection groups within the distributed network of storage elements; wherein at least some of the other protection group members of a first of the at least two separate protection groups and at least some of the other protection group members of the second of the at least two separate protection groups are not commonly assigned to the same storage element; and wherein the protection group members are used to store data from an erasure code.

42 . The method of claim 41 , wherein the erasure code has a split drive fault span including a combination of a first number of protection group members over which data is dispersed plus a second number of redundancy protection group members.

43 . The method of claim 42 , wherein no more than the second number of protection group members are contained in any hardware protection zone, wherein the hardware protection zones comprise subsets of storage elements physically separated and supported by independent power supplies and network infrastructure.

44 . The method of claim 42 , wherein each of the separate protection groups has a size including a predetermined number of protection group members, the method further comprising the step of combining two or more separate protection groups to enable storage of erasure codes having a split drive fault span which exceeds a size of a single protection group.

45 . A method of allocating RAID storage volumes across a distributed network of storage elements, the method comprising the steps of:
defining at least two logical splits on each of the storage elements; providing RAID protection for the splits of a first of the storage elements using at least two RAID groups, each of the at least two RAID groups encompassing a plurality of other RAID members stored on splits of other storage elements of the distributed network of storage elements; and distributing the other RAID members of the at least two RAID groups within the distributed network of storage elements; wherein at least some of the other RAID members of a first of the at least two RAID groups and at least some of the other RAID members of the second of the at least two RAID groups are not commonly assigned to the same storage element; and wherein the step of distributing the RAID protection within the distributed network of storage elements is implemented using a set of heuristics defining placement rules for members of the RAIDs, wherein the heuristics comprise hard rules and soft rules.

46 . The method of claim 45 , wherein the hard rules include a first hard rule prohibiting placement of two RAID members on a given storage element and a second hard rule requiring placement of RAID members in separate hardware protection zones within the distributed network of storage elements, wherein the hardware protection zones comprise subsets of storage elements physically separated and supported by independent power supplies and network infrastructure, wherein the subsets of storage elements are divided into logical subsets to support logic redundancy.

47 . The method of claim 45 , wherein the soft rules comprise a first soft rule seeking to minimize overlap between common sets of RAID members on subsets of storage elements.

48 . The method of claim 47 , wherein the soft rules are applied at a lower priority level to organize RAID groups efficiently to meet the hard rules.

49 . The method of claim 47 , wherein soft rule parameters are prioritized to maximize soft rule benefits in the final distribution.

50 . A non-transitory tangible computer readable storage medium having stored thereon a computer program for implementing a method of allocating RAID storage volumes across a distributed network of storage element, the computer program comprising a set of instructions which, when executed by a computer, cause the computer to perform a method comprising the steps of:
defining at least two logical splits on each of the storage elements; providing RAID protection for the splits of a first of the storage elements using at least two RAID groups, each of the at least two RAID groups encompassing a plurality of other RAID members stored on splits of other storage elements of the distributed network of storage elements; and distributing the other RAID members of the at least two RAID groups within the distributed network of storage elements; wherein at least some of the other RAID members of a first of the at least two RAID groups and at least some of the other RAID members of the second of the at least two RAID groups are not commonly assigned to the same storage element; and wherein the step of distributing the RAID protection within the distributed network of storage elements is implemented by defining a matrix based on a set of splits within a first hardware fault zone, and replicating the matrix into other fault zones until all RAIDs have a complete set of members, wherein the matrix is created by forming an matrix template based on a placement of members of the RAID groups such that no two RAID group members from separate RAID groups are stored on more than one storage element in common.

51 . The non-transitory tangible computer readable storage medium of claim 50 , wherein the matrix is created by reducing the matrix template based on a number of storage elements in the first hardware fault zone of the distributed network of storage elements.

52 . The non-transitory tangible computer readable storage medium of claim 50 , the method further comprising storing data within the RAID groups.

53 . The non-transitory tangible computer readable storage medium of claim 52 , the method further comprising recovering data upon failure of a storage element for each split of the failed storage element from the distributed RAID protection within the distributed network of storage elements.

54 . The non-transitory tangible computer readable storage medium of claim 53 , wherein the step of recovering data is performed in parallel for each of the splits of the failed storage element to accelerate data recovery.

55 . The non-transitory tangible computer readable storage medium of claim 50 , wherein the step of distributing the RAID protection within the distributed network of storage elements is implemented using a set of hard rules and soft rules defining placement rules for members of the RAIDs, and wherein RAID storage elements of two or more partitions occupying the same physical fault zone are redistributed in the fault zone to further optimize the distributions in accordance to the soft and hard rule requirements.

56 . The non-transitory tangible computer readable storage medium of claim 50 , wherein the step of distributing the other RAID members of the at least two RAID groups within the distributed network of storage elements comprises performing a simulation to generate at least one RAID group distribution map.

57 . A non-transitory tangible computer readable storage me

Assignees

• Emc Ip Holding Co Llc

Inventors

• Robins Edward S
• Granlund Kevin
• Pai Seema
• Malkevich Evgeny
• Ives Stephen Richard
• Raz Roii
• Bejerano Barak

Classifications

• G06F3/0644

  Management of space entities, e.g. partitions, extents, pools · CPC title

• G06F3/067

  Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS] · CPC title

• G06F3/0631 Primary

  by allocating resources to storage systems · CPC title

• G06F3/0604

  Improving or facilitating administration, e.g. storage management · CPC title

• G06F3/0619

  in relation to data integrity, e.g. data losses, bit errors · CPC title