MODULE – 4
VIRTUALIZED DATA
CENTER – STORAGE
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 1
Module 4: Virtualized Data Center – Storage
Upon completion of this module, you should be able to:
• Explain storage virtualization and its implementation
• Explain virtual machine storage options
• Describe block and file level storage virtualization
• Describe virtual provisioning and automated storage tiering
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 2
Module 4: Virtualized Data Center –
Storage
Lesson 1: Storage Virtualization Overview
Topics covered in this lesson:
• Key benefits of storage virtualization
• Implementation of storage virtualization at compute, network,
and storage layers
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 3
Storage Virtualization
Storage virtualization
It is the process of masking the underlying complexity of physical
storage resources and presenting the logical view of these
resources to compute systems.
• Logical to physical storage mapping is performed by the
virtualization layer
• The virtualization layer abstracts the identity of physical storage
devices
Creates a storage pool from multiple, heterogeneous storage arrays
• Virtual volumes are created from the storage pools and are
assigned to the computing system
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 4
Benefits of Storage Virtualization
• Adds or removes storage without any downtime
• Increases storage utilization thereby reducing TCO
• Provides non-disruptive data migration between storage devices
• Supports heterogeneous, multi-vendor storage platforms
• Simplifies storage management
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 5
Storage Virtualization at Different Layers
Layers Examples
Compute • Storage provisioning for VMs
• Block-level virtualization
Network
• File-level virtualization
• Virtual Provisioning
Storage
• Automated Storage Tiering
Storage virtualization may be implemented at compute, network, and storage
layers. At the compute layer, hypervisor allocates storage space for VMs without
exposing the complexity of the physical storage. Block and file level
virtualization are network-based virtualization techniques which embed
intelligence of virtualizing storage resources at network layer.
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 6
Module 4: Virtualized Data Center –
Storage
Lesson 2: Virtual Machine Storage
Topics covered in this lesson:
• Virtual machine storage options
• Virtual machine storage considerations
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 7
Storage for Virtual Machines
Compute 1 Compute 2
• VMs are stored as set of files on
storage space available to VM 4
VM 3
hypervisor
• ‘Virtual disk file’ represents a
virtual disk used by a VM to store Virtual disk Virtual disk Virtual disk Virtual disk
its data file file file file
NFS
• Size of virtual disk file represents VMFS
storage space allocated to virtual
disk FC SAN IP Network
• VMs remain unaware of
Total space available to the
hypervisor
Underlying storage technologies
FC Storage iSCSI NAS
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 8
File System for Managing VM Files
• Hypervisor uses two file systems to manage the VM files
Hypervisor’s native file system called Virtual Machine File System
(VMFS)
Network File System (NFS) such as NAS file system
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 9
Virtual Machine File System (VMFS)
Compute 1 Compute 2
• Hypervisor’s native file system to
manage VM files
• Cluster File System
Can be accessed by multiple compute
systems simultaneously
Provides on-disk locking
• Uses a VMFS volume to store VM
files
Virtual disk Virtual disk Virtual disk Virtual disk
VMFS Volume
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 10
Dynamic Expansion of VMFS
Expand VMFS on the existing
• VMFS can be dynamically expanded volume
without disrupting running VMs
VMFS
Methods to expand VMFS
Add a LUN to the existing VMFS
• Expand VMFS dynamically on the volume volume
partition on which it is located VMFS
LUN VMFS Volume
• Add one or more LUNs to the source VMFS Volume
volume
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 11
Raw Device Mapping
• Enables VM to directly access LUNs
in a storage system VM1 VM2
• Contains a symbolic link on VMFS
volume to the LUN
Acts as a proxy that allows direct VM Mapping
Content File
access to a LUN
VMFS Volume
Benefits VM
Content
• Provides solution when huge volume of data on LUN on Physical
LUN is not practical to move onto virtual disk Storage System
• Enables clustering the VM with physical machine
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 12
Network File System
Compute 1 Compute 2
• Hypervisor uses NFS protocol to
access NAS file system
• NFS volumes are created on NAS
device
Provide storage to VM
Accessed by multiple compute systems
simultaneously
NFS Volume
Virtual disk Virtual disk Virtual disk Virtual disk
NAS System
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 13
Module 4: Virtualized Data Center –
Storage
Lesson 3: Block-level and File-level Virtualization
Topics covered in this lesson:
• Block-level storage virtualization
• File-level storage virtualization
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 14
Block-level and File-level Virtualization –
Overview
• Network-based virtualization embeds storage virtualization
intelligence at the network layer
• Provides ability to
Pool heterogeneous storage resources
Perform non-disruptive data migration
Manage a pool of storage resources from a single management interface
• Network-based storage virtualization is applied at
Block-level (SAN)
File-level (NAS)
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 15
Block-level Storage Virtualization
Compute
• Creates an abstraction layer at
SAN, between physical storage
resources and volumes presented
to compute
Virtual volume
• Uses virtualization appliance to
perform mapping operation
• Makes underlying storage
infrastructure transparent to SAN
compute Virtualization
• Enables significant cost and Appliance
resource optimization
Heterogeneous Storage Arrays
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 16
Physical to Virtual Volume Mapping
Compute Compute
Virtualization
Appliance
Virtual Virtual
volume volume
Extent Extent Extent Extent
Storage Volume Storage Volume Storage Volume Storage Volume
Storage Storage
Array Array
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 17
File-level Storage Virtualization Clients
• Provides an abstraction in the NAS/File
servers environment
Eliminates dependencies between the file
Virtualization
and its location Appliance
• Enables movement of files between IP
NAS systems without impacting client IP Network
access
• Provides opportunities to optimize
storage utilization
• Implemented using global namespace
Multi-vendor NAS Systems
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 18
File-level Storage Virtualization – Global
Namespace
• Enables clients to access files using logical names which are
independent of the actual physical location
• Maps logical path of a file to the physical path names
• Simplifies access to files
Clients no longer need to have multiple mount points to access data
located on different NAS devices
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 19
Module 4: Virtualized Data Center –
Storage
Lesson 4: Virtual Provisioning and Automated Storage Tiering
Topics covered in this lesson:
• Virtual provisioning and its benefits
• Thin LUN and Thin Pool
• Virtual Provisioning for virtual disks
• Automated Storage Tiering
• Sub-LUN Tiering and Cache-Tiering
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 20
Virtual Provisioning (Thin Provisioning)
Virtual Provisioning (Thin Provisioning)
Compute Systems
It is the ability to present a LUN to a
compute system with more capacity than
what is physically allocated to the LUN.
10 TB 10 TB 10 TB
• Capacity-on-demand from a shared storage Thin LUN Thin LUN Thin LUN
pool, called Thin pool Compute
Reported
Physical storage is allocated only when the compute Capacity
3 TB 4 TB 3 TB
requires it Allocated Allocated Allocated
Provisioning decisions not bound by currently
available storage
• May be implemented at Disk Drives
Storage layer
Compute layer – virtual Provisioning for virtual disk
Shared Storage pool
(Thin Pool)
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 21
These challenges are addressed by Virtual Provisioning.
• Virtual Provisioning is the ability to present a logical unit (Thin
LUN) to a compute system, with more capacity than what is
physically allocated to the LUN on the storage array.
• Physical storage is allocated to the application “on-demand”
from a shared pool of physical capacity.
• This provides more efficient utilization of storage by reducing
the amount of allocated, but unused physical storage. You will
learn about Thin LUN and Thin pool later in this lesson.
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 22
Traditional Provisioning vs. Virtual Provisioning
150 GB
Available
Capacity
1500 GB 1650 GB
or 800 GB or
1.5 TB 550 GB 1.65 TB
Allocated Available
600 GB 500 GB
Allocated Unused Capacity
Unused
400 GB
500 GB Capacity
Capacity
Allocated
Allocated
Unused
Unused
Capacity
Capacity
350 GB 350 GB
100 GB 200 GB
Actual data
100 GB 50 GB 200 GB Actual data
Data 50 GB Data Data Allocated Allocated Allocated
Thin Thin Thin Storage System
LUN 1 LUN 2 LUN 3 Storage System
LUN 1 LUN 2 LUN 3 2 TB
500 GB 550 GB 800 GB 2 TB
Traditional Provisioning Virtual Provisioning
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 23
Thin LUN
• Logical device where the physical storage need not be
completely allocated at the time of creation
• Seen by the operating system as a traditional LUN
• Physical storage is allocated to the Thin LUN from the Thin pool
• Minimum amount of physical storage allocated at a time to a
Thin LUN from a Thin Pool is called Thin LUN Extent
• Best suited for environments, where space efficiency is
paramount
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 24
Thin Pool
Thin Pool
• Collection of physical drives that
provide the actual physical storage used
by Thin LUNs
• Multiple pools may be created within a
storage array
• Can be expanded dynamically
Drives can be added to a Thin pool while
pool is being used in production
• Allocated capacity is reclaimed by the
pool when Thin LUNs are destroyed Additional Disk Drives
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 25
Thin Pool Rebalancing
• Thin pool rebalancing is a technique
that provides the ability to
automatically rebalance allocated
extents on physical disk drives over the
entire pool when new disk drives are
added to the pool. Thin pool
rebalancing restripes data across all the
disk drives( both existing and new disk
drives) in the thin pool. This enables
spreading out the data equally on all
the physical disk drives within the Thin
pool, ensuring that the used capacity of
each disk drive is uniform across the
pool.
•
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 26
Virtual Provisioning at Compute
• Hypervisor performs virtual
provisioning to create virtual
disks for VMs
Virtual machine sees full logical Hypervisor
disk size at all times
• Hypervisor allocates storage
space to the virtual disk only Thin Thick Thin
when VM requires storage space 10GB
20GB 40GB
Eliminates the need to 20GB
overprovision virtual disks Virtual Disks
100GB
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 27
Virtual Provisioning Benefits
• Reduces administrative overhead
• Improves capacity utilization
• Reduces cost
• Reduces downtime
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 28
Virtual Provisioning Best Practices
• Drives in Thin pool should have same RPM
• Drives in the Thin pool should be of same size
• Provision Thin LUNs for applications that can tolerate some
variation in performance
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 29
Storage Tiering
Storage Tiering
Establishing a hierarchy of storage type, and identifying the candidate data
to relocate them to the appropriate storage type to meet service level
requirements at a minimal cost.
• Each tier is optimized for a specific characteristic, such as
performance, availability, or cost
• Efficient storage tiering requires implementation of policies
Policies may be based on parameters such as file type, frequency of
access etc.
• Storage Tiering Implementation
Manual storage tiering
Automated storage tiering
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 30
Automated Storage Tiering
• Automates the storage tiering process
• Enables non-disruptive data movement between tiers
• Improves application performance at the same cost or provides
the same performance at a lower cost
• Configures data movement
Within a storage array (Intra-array)
Between storage arrays (Inter-array)
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 31
Automated Storage Tiering – Intra Array
• Automates the storage tiering process within
array Faster,
• Enables efficient use of Solid-state drives Application
performance
SSD
(SSDs) and SATA drive technologies
Moving active data to high performance SSD tier
and inactive data to higher capacity lower Fibre
Channel
performance SATA drives tier
• Performs data movements between tiers at
Lower storage
sub-LUN level costs and less
SATA
energy
• Employs cache tiering to improve application
performance further
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 32
Sub-LUN Tiering
• By using sub-LUN tiering, a LUN is broken down into smaller
segments and tiered at that level
• Solid-state drives (SSD)
5% Active SSD
95% Inactive HDD
Storage pool
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 33
Automated Storage Tiering – Building Blocks
Building Description
Blocks
Combination of drive technologies (SSD, FC, or SATA) and
Storage Type
RAID protection
Storage groups Logical collection of LUNs that are managed together
Manage data placement and movement across Storage
Policies Types to achieve service levels for one or more Storage
Storage Type
Groups Storage Groups
Policies
Type 1 Platinum
200 GB SSD
RAID 5 (3+1) 25%
Exchange_SG
50%
Type 2 25%
146 GB 15K FC
RAID 1 Silver Oracle_SG
Type 3 25%
1 TB SATA 100%
RAID 6 (14+2)
Dev_SG
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 34
Cache Tiering
• Creates a large capacity secondary
cache using SSDs Hypervisor Hypervisor Hypervisor
• Enables tiering between DRAM cache,
SSDs drives (secondary cache) Array controller
• Most reads are now served directly DRAM Cache Tier 1
from high performance tiered cache Tiered
Cache
SSD Tier 2
Benefits
• Provides excellent performance benefit
during peak workload
• Non-disruptive and transparent to
applications
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 35
Module 4: Summary
Key points covered in this module:
• Storage virtualization and its benefits
• Storage for Virtual Machines
• Network-based storage virtualization
• Virtual provisioning and its benefits
• Automated storage tiering
Copyright © 2011 EMC Corporation. All Rights Reserved. Virtualized Data Center - Storage 36