©2011 The Enterprise Strategy Group
Enterprise Strategy Group | Getting to the bigger truth.T M
Overcoming the Storage Challenges Associated With Server Virtualization
Terri McClure October 6, 2011
What You Will Learn Today
• The current state of server virtualization deployments
• The storage challenges associated with each stage of server virtualization maturity
• The impact server virtualization will have on storage management and tools
• How scale out storage solutions address the key challenges associated with supporting virtual server environments
© 2011 Enterprise Strategy Group
©2011 The Enterprise Strategy Group
Enterprise Strategy Group | Getting to the bigger truth.T M
How Far Along is Server Virtualization?
4
The Four Dimensions of ESG’s Server Virtualization Maturity Model
© 2011 Enterprise Strategy Group
Source: ESG Research Report: The Evolution of Server Virtualization, November 2010.
Dimension: Scope of deployment
Metric: Percent of servers that have been virtualized
Dimension: Virtual production ratio
Metric: Percent of VMs in production
Dimension: EfficiencyMetric: Consolidation
ratio
Dimension: Workload
penetrationMetric: Deployment
across multiple workloads
5
ESG Server Virtualization Maturity Model Segmentation
© 2011 Enterprise Strategy Group
Source: ESG Research Report: The Evolution of Server Virtualization, November 2010.
Gro
up 1 Basic
Low percent of physical servers virtualized (typically less than 20%)Low percent of VMs deployed in production (12% average)Low virtual-to-physical consolidation ratio (typically less than 5)Deployment across limited and/or basic workloads (13% have deployed mission-critical production workloads)
Gro
up 2 Progressing
Medium percent of physical servers virtualized (typically 10% - 30%)Medium percent of VMs deployed in production (35% average)Medium virtual-to-physical consolidation ratio (typically 5 - 10)Deployment across more workloads, but mostly tier-2 (37% have deployed mission-critical production workloads)
Gro
up 3 Advanced
High percent of physical servers virtualized (typically 40% +)High percent of VMs deployed in production (67% average)High virtual-to-physical consolidation ratio (typically more than 10)Deployment across many/tier-1 workloads (75% have deployed mission-critical production workloads)
Length of Server Virtualization Deployment
© 2011 Enterprise Strategy Group 6
Less than 12 months; 9%
1 to 3 years; 57%
4 to 5 years; 18%
More than 5 years; 16%
For how long has your organization had server virtualization technology deployed? (Percent of respondents, N=463)
Percentage of Servers That Have Been Virtualized
© 2011 Enterprise Strategy Group 7
Less than 10% of servers
10% to 20% of servers
21% to 30% of servers
31% to 40% of servers
41% to 50% of servers
More than 50% of servers
Don’t know0%
5%
10%
15%
20%
25%
30%
35%
40%
12%
21%
25%
15%
11%14%
1%2%
9%11%
17%20%
38%
3%
Of all the potential x86 servers in your organization that can be virtualized, approx-imately what percentage of these systems have actually been virtualized to date? How do you expect this to change over the next 24 months? (Percent of respon-
dents, N=463)
Today 24 months from now
Percentage of VMs Deployed in Production
© 2011 Enterprise Strategy Group 8
Today (N=440) 24 months from now (N=434)0%
10%
20%
30%
40%
50%
60%
70%
39%
58%
Approximately what percentage of your total virtual machines are currently running in a production environment? How do you expect this to change in the
next 24 months? (Mean response)
9
Storage Infrastructure Challenges Related to Server Virtualization Usage
© 2010 Enterprise Strategy Group
Source: ESG Research Report: The Evolution of Server Virtualization, November 2010.
We have not encountered any challenges
Need different tools to manage storage environment and virtual server environment
Difficulty backing up and/or recovering virtual machines
Sizing capacity required to support virtual server environment
Sizing performance requirements (i.e., IOPS) required to support virtual server environment
General lack of information or best practices
Lack of integration between storage system and virtual machine replication capabilities
Need to conduct additional interoperability testing and qualification
Impact on overall volume of storage capacity
Performance problems
Security concerns
Operational cost of new storage infrastructure
Scalability problems
Capital cost of new storage infrastructure
0% 5% 10% 15% 20% 25% 30% 35% 40%
6%
12%
15%
17%
17%
17%
17%
19%
21%
21%
24%
24%
25%
36%
From a storage infrastructure perspective, which of the following would you consider to be significant challenges related to your organization’s server virtualization usage? (Percent of respondents, N=190, multiple responses
accepted)
© 2011 Enterprise Strategy Group
We have not encountered any challenges
Lack of integration between storage system and virtual machine replication capabil-ities
Security concerns
Scalability problems
Sizing performance requirements (i.e., IOPS) required to support virtual server en-vironment
General lack of information or best practices
Need different tools to manage storage environment and virtual server envi-ronment
Performance problems
Sizing capacity required to support virtual server environment
Need to conduct additional interoperability testing and qualification
Operational cost of new storage infrastructure
Difficulty backing up and/or recovering virtual machines
Impact on overall volume of storage capacity
Capital cost of new storage infrastructure
0% 5% 10% 15% 20% 25% 30% 35% 40%
3%
21%
28%
34%
31%
24%
14%
17%
17%
17%
21%
7%
21%
38%
5%
21%
30%
30%
16%
17%
10%
24%
17%
21%
27%
14%
19%
38%
13%
9%
11%
11%
15%
15%
17%
17%
20%
22%
22%
26%
28%
37%
From a storage infrastructure perspective, which of the following would you consider to be significant challenges related to your organization’s server virtualization usage?
(Percent of respondents, N=190, multiple responses accepted)
Advanced (N=46)
Progressing (N=111)
Basic (N=33)
Storage-Specific Challenges Related to Server Virtualization Environment, by Maturity Phase
11
Basi
c Le
vel CA PEX of new
storage
Scalab ility
Sizing perform ance requirem ents
Security
Lack of inform ation and/or best practices
CAPEX of new storageScalabilitySizing performance requirementsSecurity Lack of information and/or best practices
Prog
ress
ing
Leve
l CA PEX of new storage
Scalab ility
Security
O PEX of new storage
Interoperability testing & qualification
Integration of storage and VM replication
CAPEX of new storage Scalability Security OPEX of new storageInteroperability testing & qualification Integration of storage and VM replication
Adva
nced
Lev
el
CA PEX of new storage
O verall storage volum e
Backup and/or recovery of VM s
Interoperability testing & qualification
O PEX of new storage
CAPEX of new storage Overall storage volumeBackup and/or recovery of VMsInteroperability testing & qualificationOPEX of new storage
Top Five Storage Challenges, by Maturity Phase
Source: ESG Research Report, The Evolution of Server Virtualization, November 2010.
Impact to Storage Management Process & Tools
Basi
c Le
vel
Prog
ress
ing
Leve
l
Adva
nced
Lev
el
54%Significant
Some
Little or none
39%
7%
35%
59%
2%
24%
52%
15%
(Migrate most or all storage management functions to a virtualization software
platform)
(Migrate some storage management functions to a virtualization software
platform, but continue to use existing tools)
(Use storage management functions in the virtualization software platform to a
minor extent, will not fundamentally change existing tools)
© 2011 Enterprise Strategy Group
Storage-specific Factors That Would Enable More Widespread Usage of Server Virtualization
13
QoS tools to allocate storage resources appropriately
Improved multi-tenancy capabilities (i.e., ability for multiple users and/or applications to securely share storage resources)
Need better integration with array-based replication technology
New consumption-based pricing model (i.e., only pay for storage resources being used)
Improved storage management capabilities optimized for virtual envi-ronments
More integration between storage and virtualization orchestration, man-agement, operations, etc.
Need backup technology designed for server virtualization
More scalable storage infrastructure to support rapid virtual machine growth
Better storage migration tools
Increased use of storage virtualization
Faster storage provisioning
Additional training for IT staff
0% 5% 10% 15% 20% 25% 30% 35%
13%
16%
16%
18%
18%
20%
24%
25%
25%
25%
28%
33%
From a storage infrastructure perspective, which of the following developments need to take place in order to enable more widespread server virtualization usage? (Percent of respon-
dents, N=190, multiple responses accepted)
© 2011 Enterprise Strategy Group
Netting it Out
• Server virtualization stresses the surrounding infrastructure
• Creates new demands on the storage infrastructure• Changes the storage management paradigm• Requires a more scalable and cost effective storage
infrastructure
© 2011 Enterprise Strategy Group 14
©2011 The Enterprise Strategy Group
Enterprise Strategy Group | Getting to the bigger truth.T M
Scale-out StorageMeeting the needs of virtual server environments with a virtual storage environment
ESG Research Findings
© 2011 Enterprise Strategy Group 16
• More than 1 in 4 users surveyed report that supporting server virtualization initiatives is the IT initiative that will have the greatest impact on storage spending (#1 response)
• Improved scalability, performance, faster provisioning and lower infrastructure and management costs drive scale-out adoption in enterprise IT
• Scale-out usage is much more common among organizations in which storage is viewed as strategic to IT and the business
• Scale-out users more likely to have a sufficient number of administrators to manage their storage and are more satisfied with utilization rates
© 2010 Enterprise Strategy Group 17
Current Challenges Scaling Storage Environments
Difficulty maintaining up-to-date information on current in-frastructure inventory and configurations
Running out of physical space
Capacity balancing
Management, optimization and automation of data placement
Need to support growing virtual server environments
Data migration
Hardware costs
Data protection
Rapid growth and management of unstructured data
Operational costs
0% 5% 10% 15% 20% 25% 30% 35%
11%
15%
17%
17%
20%
21%
28%
32%
20%
31%
4%
4%
5%
6%
8%
8%
9%
9%
11%
11%
Organizations’ biggest challenges (all and primary) in terms of scaling storage en-vironment to support its applications? (Percent of respondents, N=306, top ten re-
sponses shown)
Primary stor-age challenge
All storage challenges
© 2010 Enterprise Strategy Group 18
Most Common Reason For Adding New Storage Systems
Need additional capacity and we can’t expand existing system(s);
25%
Need improved I/O performance; 7%
Need improved throughput per-formance; 10%
Need improved reliability; 9%Need improved availability; 8%
Technology refresh; 20%
Need additional feature/functional-ity; 9%
Consolidation of existing storage devices; 10%
Other; 1% Don’t know; 1%
When your organization purchases and deploys new storage systems to accom-modate application data growth, what would you say is the most common reason
why you have added those new systems? (Percent of respondents, N=306)
© 2010 Enterprise Strategy Group 19
Length of Implementation Cycle for New Storage Systems
Less than 3 months; 17%
3 to 6 months; 29%
7 to 9 months; 22%
10 to 12 months; 21%
13 to 18 months; 8%
19 to 24 months; 2%
More than 24 months; 1% Don’t know; 1%
On average, approximately how long is the implementation cycle – from planning to production – to add new storage systems? (Percent of respon-
dents, N=306)
Scale-Up vs. Scale-Out StorageIn
tern
al n
etw
ork
(bus
or
hig
h-sp
eed
mem
ory)
Storage system
“Scale-Up” Storage: Scales by adding resources within a single system, connected by an
internal network
“Scale-Out” Storage: Scales by adding nodes -
connected by an external network – that work together as a single
system
Multiple nodes acting as one storage system
External network
System resources (e.g., processors,
memory, host interfaces, disks)
System resources (e.g., processors,
memory, host interfaces, disks)
System resources (e.g., processors,
memory, host interfaces, disks)
Additional system resources
Disk resources
Disk resources
Disk resources
Disk resources
Disk resources
Disk resources
Additional disk
resources
System resources (e.g., processors,
memory, host interfaces)
21
Why Scale-Out Storage?
• Management efficiencies reduce the headcount required to provision, tune and manage storage
• Faster provisioning increases ability to respond to a dynamic virtual server environment
• Granular, or just-in-time, scalability allows IT to ride the commodity curves associated with declining disk and processor pricing, reducing CAPEX
• Better utilization means fewer systems to manage and less wasted floor space, power and cooling
• Scaling performance with capacity allows for greater capacity, more throughput and broader application support
Use of Scale-out Storage Architectures, All Respondents
22
Already using scale-out storage; 18%
Plan to use within the next 12 months; 26%
Plan to use within the next 24 months; 14%
No plans at this time, but we are interested in this technology; 26%
No plans or interest at this time; 9%
Not familiar with scale-out storage technology; 4%
Don’t know; 3%
Based on the previous definition, please describe your organization’s usage of or interest in scale-out storage architectures. (Percent of respondents, N=306)
© 2011 Enterprise Strategy Group
© 2010 Enterprise Strategy Group 23
Length of Implementation Cycle for New Storage Systems, by Scale-Out Storage Usage
Less than 3 months
3 to 6 months 7 to 9 months 10 to 12 months
13 to 18 months
19 to 24 months
More than 24 months
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
38%41%
5%
13%
4%
0% 0%
13%
26% 25%23%
9%
2% 1%
Length of new storage system implementation cycle, by scale-out storage usage (Percent of respondents)
Already using scale-out storage (N=56) Not using scale-out storage (N=242)
Most Important Considerations in Deploying Scale-out Storage, Early Adopters
24
Need to support specific applications
Improved data management
Improved storage hardware utilization
Reduced operating expenditures
Improved data availability
Improved performance (I/Os)
Easier to manage
Lower cost of infrastructure
Faster storage provisioning times
Improved performance (throughput)
Improved scalability
0% 10% 20% 30% 40% 50% 60%
16%
23%
25%
27%
27%
32%
36%
38%
38%
39%
50%
Which of the following considerations were most important in your organiza-tion’s decision to deploy scale-out storage? (Percent of respondents, N=56,
multiple responses accepted)
© 2011 Enterprise Strategy Group
Benefits of Deploying Scale-out Storage, Early Adopters
25
Reduced training time/costs
Reduced capital expenditures
Ability to manage more storage capacity with fewer administrator resources
Improved data management
Ability to more effectively support specific applications
Reduced operational expenditures
Improved storage hardware utilization
Faster deployments/provisioning times
Improved data availability
Improved performance (throughput)
Improved performance (I/Os)
Improved scalability
0% 10% 20% 30% 40% 50% 60%
14%
18%
25%
27%
27%
27%
29%
34%
36%
39%
45%
57%
Which of the following benefits has your organization realized as the result of deploying scale-out storage? (Percent of respondents, N=56, multiple responses accepted)
© 2011 Enterprise Strategy Group
Types of Data Organizations Expect to Store on Scale-Out Storage Solution, Planned Adopters/Interested
26
Archive data (i.e., data that has been removed from production systems due to factors such as age of the data, infre-
quency of access, and/or regulatory compliance requirements)
Primary application data (i.e., active data)
Secondary data (i.e., backups/copies of primary data)
0% 10% 20% 30% 40% 50% 60% 70%
33%
47%
63%
Please indicate which of the following types of data you expect to store on your organization’s scale-out storage system(s). (Percent of
respondents, N=202, multiple responses accepted)
© 2011 Enterprise Strategy Group
Types of Data Stored on Scale-out Storage Solution, Early Adopters
27
Archive data (i.e., data that has been removed from production systems due to factors such as age of the data, infrequency of access, and/or regulatory com-
pliance requirements)
Secondary data (i.e., backups/copies of primary data)
Primary application data (i.e., active data)
0% 10% 20% 30% 40% 50% 60% 70% 80% 90%
41%
61%
82%
Please indicate which of the following types of data is stored on your organiza-tion’s scale-out storage system(s). (Percent of respondents, N=56, multiple re-
sponses accepted)
© 2011 Enterprise Strategy Group
Adoption Trends Summary
• Scale-out storage use is on the rise• The promise and benefits are real, not marketing spin • Because of these benefits we see “usage creep”• Storage is becoming more of an IT infrastructure
decision, less tied to single applications • ESG expects 80% of external networked storage
shipment revenue in 2015 to be from scale-out systems
© 2011 Enterprise Strategy Group 28
The Bigger Truth
• Server virtualization impacts all adjacent technology areas The storage environment needs to be as virtual as the
server environment
• Scale-out storage is well aligned to supporting server virtualization It virtualizes the underlying storage hardware
infrastructure; scales quickly and efficiently; helps drive down operational costs
• Scale-out storage can help ease users journey along the path to broadening server virtualization initiatives
© 2011 Enterprise Strategy Group 29
Getting to the bigger truth.Getting to the bigger truth.
© 2009 Enterprise Strategy Group
TM
Thank YouFor more information, please contactTerri McClure, Senior Analyst508.381.5162 | [email protected]