Date post: | 16-Dec-2015 |
Category: |
Documents |
Upload: | owen-farmer |
View: | 215 times |
Download: | 3 times |
The StorSimple Approach to Solving Issues Related to Growing Data Trends Marc Farley
DCIM-B310
Data growth trends
From 4.4 today to 44 trillion GB in 2020
2013 2014 2015 2016 2017 2018 2019 20200
10
20
30
40
50
Worldwide Data Growth
Source: EMC Digital Universe with Research & Analysis by IDC, 2014
In 2014, the digital universe will equal 1.7 megabytes a minute for every person on Earth
1.7 MB x xSource: EMC Digital Universe with Research & Analysis by IDC, 2014
In 2013: 66% of the digital universe was created/captured by consumers & workers, but corporations had liability for 85% of it
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%Total digital universe
85% corporations had liability for
66% created by consumers & workers
78% of the data corporations were liable for was created or captured by consumers & workers
Sources of data & corporate liabilities for it
Source: EMC Digital Universe with Research & Analysis by IDC, 2014
Source: EMC Digital Universe with Research & Analysis by
IDC, 2014
The number of gigabytes per IT professional will grow by a factor of eight between now and 2020
Storage and data management have to change*
* (again)
Adapting to 40% compound data growthMatch data usage requirements to resource costsClasses of storage and their costsApplication workload requirementsLife cycle data management requirements
Increase efficiency of storage and staffEliminate special purpose storageConsolidate management effortsAutomate as much as possible
Maintain or improve recoverabilityRTOs will increase if they are based on amount of data storedDR testing is imperative for sound business continuity
Managing storage costs with 40% compound data growth rates
Lower the average cost of storageUse multiple storage levels
$$$ highest cost reserved for qualified workloads
$$ mid-level cost & performance
$ lowest cost, where most data belongs
Cloud storage
HDDs (disk drives)
SSDs (solid state drives)
Online storage technology price/performance levels
$$$
$$
$
High cost per TBLow cost per IOP
Lowest cost per TB, but impractical for IOPs
Balanced between IOPs & TB
Capacity costs
IOPs costs
$
$$$
n/a
Automated data tiering
SSD
HDD
Cloud storage
Transactional data
Backup, archive and dormant data
Business applications, office
IOPsworkloads
Casual workloads
Dormant workloads
Data life cycle requirements
SSD
Cloud
New writes need fast acknowledgements
New data needs off-site protection
Newdata HDD
“Oops” data copies in local snapshots
Life cycle of unstructured data
SSD
Cloud
Access frequency HDD
Day 1
Day 22
Day 45
Day 70
Eliminate special purpose storageAvoid multiplying data growth costs
Primarystorage
Backupstorage
Archivestorage
Off-sitestorage
Primary
Backup
Archive
Off-site
Primary data stored across all tiers
Cloud
HDD
SSD
Primary
Backup
Archive
Off-site
SSD
HDD
Backup, archive & off-site data in the cloud
SSD
Cloud
HDD
Primary
Backup
Archive
Off-site
SSD
HDD
Thinly provision storageKeep free space free as long as possible
Managing storage with 40% data growth is much easier with the maximum free space available
Consolidate managementDon’t manage more than you have to
Primary
storage
Backup
storage
Archivestorage
Off-site
storage
One dashboard, with all status, operating and alert informationPrimar
y
Backup
Archive
Off-site
1
4
3
2
Multiple management points
Automate where possibleApply computer science
Tier in SSD Use SSD IOPs to increase storage efficiency
Make data portable for local and cloud tiers
Tier to HDD
Tier to cloud
Add management information that is portable
Increase the effective capacity of HDDs Electronically protect data in the cloud
Secure data that is going to the cloud
Make redundant copies in the cloud
Lowering RTOs with 40% data growth Intelligence is the only way out of this
The “pig in the pipe” approach won’t work
2nd
2nd3rd
2nd2nd
1st 1st
3rd
3rd
3rd
2nd 2nd3rd
3rd
Prioritize data for recovery
Emergency preparation takes practiceThe disruption of DR testing must be diminished
Plan
DoCheck
Act
Plan
DoCheck
Act
Managing data growth with StorSimple hybrid cloud storage
StorSimple design points
Comparing hybrid storage with hybrid cloud storageTopology viewFeature breadth and depthReduced costs of storage and management
Making the most out of SSD IOPsLinear layerPrimary deduplication Transformation to objects and metadata
Data protection and recoveryCloud snapshotsThin recoveryDR Testing
StorSimple topologyiSCSI SAN front end with an object storage back end
Hybrid storage with a big differenceCloud storage integration
(Nothing)(Nothing)
Other hybrid storage with SSDs
+ HDDsSSD
HDD
SSD
HDD
Cloud storage
StorSimple’s hybrid storage: SSDs + HDDs + cloud
IOPs Office
Data protectionArchiveDormant
IOPs Office
Data protection
Archive Dormant
Cost comparisonsHybrid storage vs hybrid cloud storage
SSD
HDD
SSD
HDD
Cloud storage
Average cost of storageManagement costsUp to 90% less
Hybrid storage SSDs + HDDs
Hybrid storage SSDs + HDDs + Cloud
New writes written to the linear tier
SSD
Cloud
New writes need fast acknowledgements No deduplication Like cache, but non-volatile
Newdata HDD
Raw data in SSD queue #1
Linear tier to dedupe tierThe creation of objects and metadata
New data enters the linear tier (SSD)
Oldest data in linear tier “spills over”
Inline deduplication process creates objects and metadata
Data objects placed in dedupe tier (SSD)
Newdata
Metadata stored in NV-RAM or
SSD
Linear
Dedupe
SSDs split into two tiers
Uses the IOPs
power of SSDs to increase
utilization
Primary dedupe counteracts data growthSmall capacity multiples are big cost savings
A B C A D B D C
A B C D
A
Most effective inVM environmentsOffice environments
Blocks in
Encapsulated block objects out
Objects and metadataPortability, immutability, scalability and availability
SHA-256 hash, derived from contentsThe object handleBuilt –in checksum
Objects are the same, on premisesor in the cloud.
No import/export compute “fees”
Capacity
Nothing scales like object storage, that’s why cloud storage services use it
Online on-premises Online in the cloudLocation independence
More data reduction gains for HDDsObjects that can be compressed, are
New objects enters the dedupe tier (SSD)
Oldest objects spill over
Inline compression process works on compressible data
Data objects placed in compressed tier (HDD)
Deduped objects from
linear tier
From on-premises to the cloudIt’s not quite what you’d expect…
New objects enters the compressed tier (HDD)
Dormant objects spill over
Pointers to dormant objects are changed to reflect their cloud location
Objects from compressed
tier
Most objects are already in the cloud via cloud snapshots
Cloud snapshots
SSD
Cloud
New data sent to the cloud for protection
Newdata
HDD
Data is encrypted on-premises prior to transmission with customer-managed keys
Automated data protection to the cloud
On-premises and in the cloud
SSD
Cloud
Active objects on-premises and in the cloud
HDD
StorSimple is not a cloud cache, nor a gateway
Dormant objects only in the cloud
Workload profiles and capacitySmall percentage of data supporting most of the work
Inactive data
Backups
Archives
Disaster Recovery
IOPs
Office
BackupArchiveDormant
Thin recovery reduces RTOsRestore that small percentage of working data first
Inactive data
Backups
Archives
Disaster Recovery
Top priorityapplications
2nd priority applications
Dormant
2n
d
2n
d
3r
d
2nd2nd
1st 1st
3rd
3rd
3rd
2nd 2nd3rd
3rd
Prioritize data for recovery
Steps to a thin recoveryThe importance of portable metadata
The recovery
StorSimple system
• Credentials and keys
Download metadata
• 3% of last data image
Applications can “see” their data
• Download on access
Thin recovery and non-disruptive DR testingUse another StorSimple system at another site or a secondary StorSimple system
Thin recovery can be easily tested and should be
Know your capabilities and have expectations based on actual test experiences
Advantages of Microsoft Azure Storage3 kinds of data replication, local, zone and geo
Data integrity encoding and background scrubbing
Availability of datacenters
DCIM-B250 StorSimple Solutions with Veeam in Production Environments Tuesday, May 13 10:15 AM - 11:30 AM Room: Hilton L2 Ballrm A
Related content
DCIM-B357 StorSimple: Enabling Microsoft Azure Cloud Storage for Enterprise Workloads Wednesday, May 14 1:30 PM - 2:45 PM Room: 351A
MSE (Microsoft Solutions Experience) Cloud and Datacenter & Infrastructure Management
I will be available: Monday, 6:30 TechExpo Welcome Reception - MSE Cloud and Datacenter
Mgmt Tuesday, 2:15 PM-4:00 MSE Cloud and Datacenter Mgmt
Come Visit Us in the Microsoft Solutions Experience!
Look for Datacenter and Infrastructure ManagementTechExpo Level 1 Hall CD
For More InformationWindows Server 2012 R2http://technet.microsoft.com/en-US/evalcenter/dn205286
Windows Server
Microsoft Azure
Microsoft Azurehttp://azure.microsoft.com/en-us/
System Center
System Center 2012 R2http://technet.microsoft.com/en-US/evalcenter/dn205295
Azure PackAzure Packhttp://www.microsoft.com/en-us/server-cloud/products/windows-azure-pack
Resources
Learning
Microsoft Certification & Training Resources
www.microsoft.com/learning
msdn
Resources for Developers
http://microsoft.com/msdn
TechNet
Resources for IT Professionals
http://microsoft.com/technet
Sessions on Demand
http://channel9.msdn.com/Events/TechEd
Complete an evaluation and enter to win!
Evaluate this session
Scan this QR code to evaluate this session.
© 2014 Microsoft Corporation. All rights reserved. Microsoft, Windows, and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other countries.The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft must respond to changing market conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information provided after the date of this presentation. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.