Secrets to Shrinking Your Storage for Microsoft Windows ...download.microsoft.com/documents/uk/enterprise/85... · company·s deduplication functionality has been implemented in over

Copyright © The Storage Writer, Inc. 2009. All Rights Reserved. The Storage Writer, Inc. Independent Storage Insights, Technical Research and Reports

Phone: (770) 643-2050 E-mail: [email protected] Web: www.thestoragewriter.com

The Storage Writer Independent Storage Insights and Reports Storage Efficiency Brief

Focus on:

Microsoft Virtua lized Environments and

NetApp Deduplication

Secrets to Shrinking Your Storage for Microsoft Windows Server 2008 Hyper-VTM

By Michele Hope

September 2009

Sponsored By: Sponsored By:

The Storage Writer

Secrets to Shrinking Your Storage for Hyper-V

Copyright © The Storage Writer, Inc. 2009. All Rights Reserved. The Storage Writer, Inc. Independent Storage Insights, Technical Research and Reports

Phone: (770) 643-2050 E-mail: [email protected] * Web: www.thestoragewriter.com

July 2008 Sponsored Sponsored By:

Table of Contents

Executive Summary ................................ ................................ ................................ ................... 1

Server Virtualization: The Promise… ................................ ................................ ......................... 2

. . . and the Struggle to Achieve It ................................ ................................ .............................. 3

Understanding Hyper-V Duplicate Data from a Storage Perspective................................ .......... 4

NetApp Deduplication: A Hidden Asset for Hyper-V Environments................................ ............. 6

Lessons in Hyper-V Savings with NetApp Deduplication................................ ............................ 9

NetApp Deduplication: How to Begin ................................ ................................ ....................... 12

The Storage Writer


Copyright © The Storage Writer, Inc. 2009. All Rights Reserved. Page 1 The Storage Writer, Inc. Storage Insights, Technical Research and Reports


Executive Summary Virtual server environments offer substantial consolidation opportunities for bloated and underutilized server and desktop infrastructures. The adoption of virtual servers and virtual desktop environments has led organizations to transform how they deploy, provision, test and develop applications in order to gain the most value from this fluid and dynamic architecture. Organizations that have deployed Microsoft virtualized environments based on Windows Server 2008 Hyper-V with integrated management via Systems Center have since been able to realize great savings in both capital costs and operating expenses surrounding their use of servers. Unfortunately, as IT teams take advantage of server virtualization to better utilize their physical server resources, they can face major issues from a storage perspective. As virtual machines (VMs) proliferate, the need for storage also increases, often at an alarming rate for organizations seeking to shrink overall IT infrastructure costs. This brief looks at the challenge of harnessing storage growth within growing Microsoft virtualized environments, based on reports from third parties, analysts and current NetApp/Hyper-V users. It also explores reports of NetApp customers shrinking the amount of storage used for Hyper-V environments by more than 50%. NetApp reports it’s quite realistic for customers to expect these kinds of savings, even going so far as to guarantee them to customers.1 Citing use of several space-efficient features, the company maintains NetApp customers can store significantly more data on each NetApp volume than they could with traditional storage. To store more data (and consume less disk space), NetApp promotes the use of efficiency technologies including NetApp FlexVol for thin provisioning, NetApp Snapshot copies (for fast, small-footprint backups), and NetApp deduplication (to eliminate duplicate copies of data). NetApp deduplication, now in use in a growing number of Hyper-V installations, is one of the big keys to savings for virtual environments. (According to NetApp, the company’s deduplication functionality has been implemented in over 37,000 NetApp systems worldwide.2) This brief explores various space-saving scenarios when using NetApp deduplication in Hyper-V environments.

The Storage Writer




Server Virtualization: The Promise… Maximizing use of physical servers and their underlying resources has become increasingly important for IT organizations faced with massive application growth and its associated “server sprawl.” In recent years, huge advances have been made by Microsoft and others to offer a viable solution to these issues. Server virtualization technology now offers organizations the option to dramatically increase server utilization by consolidating many physical servers and their applications (or application roles) onto a single physical server. In this manner, prior physical servers are able to run as “virtual servers” from a stable virtual server platform. Most IT organizations have begun implementing some form of server virtualization. In fact, the findings from one Forrester report showed 78 percent of survey respondents using virtual server technology in production environments.3 Building on its familiar Windows interface and unified management capabilities under Microsoft System Center, Microsoft’s own suite of end-to-end virtualization solutions are helping customers consolidate and virtualize everything from their servers to desktops and applications. For the purposes of this paper, we’ll often refer to IT environments that use Microsoft Windows Server 2008 Hyper-V for server virtualization as Hyper-V environments, and those using Microsoft Virtual Desktop Infrastructure as virtual desktops or VDI. Virtual servers and virtual desktops managed through Hyper-V will also often be referred to as virtual machines (VMs). The benefits of adopting Hyper-V technology include: Significantly reduced capital and operating expenses surrounding the growth, acquisition and

maintenance of a much smaller footprint of physical servers and desktop PCs. These include savings associated with the use of less physical servers, less data center space, less power and cooling and less time needed to provision and manage applications.

Better IT responsiveness to develop and test production applications. (Hyper-V’s ultra-fast provisioning capabilities allow administrators to quickly roll out virtual machines that are already equipped with “pre-built” combinations of operating systems and applications.)

Higher availability and improved disaster recovery without downtime or disruption to the production environment -- all thanks to Hyper-V’s ability to offer seamless migration of VMs from one physical server to another though features like Microsoft Windows Server 2008 R2 Live Migration.

Easier build-out and support of a viable, end-to-end virtualization platform with Microsoft System Center.

Unlike the old days where the proliferation of physical servers seemed unavoidable, Microsoft virtualization solutions give customers the option to consolidate and deploy many virtual servers and virtual desktops -- all on one underlying physical host system. Management of both physical and virtual resources is accomplished using Microsoft System Center.

The Storage Writer




. . . and the Struggle to Achieve It As organizations continue plans to expand their virtual server environment by adding more VMs and taking better advantage of Hyper-V’s high availability and data mobility features, most conclude they need to move to centralized storage and a network storage platform. (In fact, shared network storage is a necessary component when trying to use Windows Server 2008 R2 Hyper-V features like Live Migration.) Among other benefits, the use of network storage allows a growing Hyper-V environment to: More easily provision storage from a common storage capacity pool Ensure virtual machines stay up and running through use of the storage platform’s enterprise-

class data availability, reliability and replication features Reduce virtual server backup windows and boost virtual server performance by offloading much

of the burden of data protection processing from the physical server environment to the storage system.

While the use of network storage helps address many mobility, reliability and availability concerns surrounding virtual machines, there are still a few areas that remain a challenge. One of those is the amount of storage consumed by growing virtual server environments. While the use of Hyper-V VMs causes the physical server footprint to shrink dramatically in the data center, the proliferation of VMs can simultaneously cause storage capacity requirements to spiral upwards. Left unchecked, extra storage needed to support a growing Hyper-V environment can ultimately negate many of the initial gains an organization hoped to achieve. In fact, respondents to a 2008 Forrester Research survey4 showed storage growth issues in virtual server environments to be three of the top five storage challenges they currently face. Among the top five storage issues identified, issues 3, 4 and 5 reflect the on-going capacity struggle many virtual environments encounter:

1. Maintaining high performance 2. Completing backups on time 3. Efficiently managing storage capacity 4. Controlling costs 5. Provisioning storage

(Emphasis added) Why is storage becoming such an issue for virtual server environments? The following sections help shed some light on the problem.

The Storage Writer




More VMs Equal More Duplicate Data

In an article on data proliferation in virtual server environments, the CTO of GlassHouse Technologies, Jim Damoulakis, framed the overall storage issue: . . .Another important consideration is the increase of the volume of data created by virtual environments. The standard use of configuration templates for creating virtual machines results in numerous cookie-cutter VMs of a particular class that are nearly identical. Backing up and storing the myriad copies of nearly redundant data [also] consumes significant storage. . .5

As Damoulakis points out, storage capacity issues not only impact the virtual server environment’s primary storage systems. Indeed, they also become a challenge when trying to achieve optimal disk-based backup or replication of the growing virtual infrastructure -- especially where replication of very large data sets is not always feasible given the cost of bandwidth and WAN connections.

Dealing with Duplicate Data -- for Virtual Servers and Virtual Desktops

The issue of exponential storage growth for virtual server environments is not just focused on those running virtual servers. In fact, the scale in which some companies may choose to deploy virtual desktops can well bring storage capacity growth issues to a head earlier in the process. While virtual desktops may only be deployed with an initial 8-10 GB of storage capacity, companies that want to deploy hundreds (or thousands) of virtual desktops will soon see a storage impact.

Understanding Hyper-V Duplicate Data from a Storage Perspective To set the groundwork for the rest of this brief, it's important to describe how duplicate data can proliferate so easily in Hyper-V environments. This involves first defining a few concepts surrounding storage of virtual machine data.

What is a Virtual Machine?

According to Microsoft documentation, a virtual machine is essentially a “virtual computer within a physical computer, implemented in software. A virtual machine runs on a virtual machine host and emulates a complete hardware system, from processor to network card.”6 Virtual resources for each virtual machine include such items as virtual CPU, RAM, a Network Interface Card (NIC) and a virtual hard disk (VHD).

Which Files Represent a VM?

From a storage administrator's perspective, each virtual machine consists of multiple files stored in their own folder or subdirectory on the storage system. (For users of NetApp storage systems, these files would be stored on LUNs within NetApp flexible volumes.) Hyper-V offers the capability to create different types of virtual disks for each VM. One of the most common types, often chosen for optimal performance, is a fixed-size VHD. For the purposes of this brief and the assessment of NetApp deduplication functionality with Hyper-V, we’ll focus most of our discussion on the use of fixed-size VHD files. Although Microsoft has increased performance for dynamically expanding virtual hard disks with the release of Windows

The Storage Writer




Server 2008 R2, fixed-size VHD files still tend to offer the greatest opportunity for optimal performance and disk space savings when combined with NetApp deduplication technology. 7 Key files within each VM’s “subdirectory” are described in Table 1.

Table 1. Key Files per Each VM in Hyper-V.

File Description *.vhd file The *.vhd file represents the contents of a single virtual hard disk,

appearing as direct-attached storage to any virtual machine. One or more *.vhd files may be in place for each virtual machine. Contents of a virtual machine's virtual hard disk(s) typically include: Operating system data used by the virtual machine. This may

include binary files, OS patches and drivers. Applications running on the virtual machine, including application

binary files, libraries, etc. These may be called application images. Application data. Other temporary data, such as VM swap data or page files.

*.xml file The *.xml file is a configuration file associated with the virtual machine. *.vsv file The *.vsv file contains saved system state data for the virtual machine. *.bin file The *.bin file contains memory allocation data for the virtual machine.

Other file types may also be found in each VM directory. These are less relevant to this brief, so will not be described here.

How More VMs Create More Duplicate Data

Among the files that represent a VM entirely, VHD files are usually the largest. They tend to consume a majority of the disk space needed to store the VM. Based on the information in Table 1 above, which covers the primary contents of most VHDs, administrators of virtual environments can expect each VHD file to contain a lot of duplicate data. This is especially true in virtual desktop infrastructures (VDI). In fact, for VMs running the same or similar operating systems and applications, you’ll find a majority of the data in their VHDs to be identical. It’s only when you begin to look at application or temporary data stored in the VHDs that the VHDs begin to look quite different. Figure 1 shows a conceptual view of the duplicate data you might find between one VM and another.

The Storage Writer




Figure 1. Example of Duplicate Data in each VM's VHD files.

In a comparison between VM1, VM2, and VM3, each conceptual layer of data (OS, Application, etc.) is shown in a different color: OS data is green Application installations are blue Application data is grey, Other miscellaneous (or “Misc.”) data is yellow Within each layer, “D” blocks on VM2 and VM3 show duplicate data also contained in VM1, while red “U” blocks on VM2 and VM3 show data that's unique from VM1. Duplicate blocks on each layer would be considered candidates for reduction via deduplication technology. In the above example with just three VMs and no deduplication technology in use, if each VHD file is 20GB, the VHD files for the three VMs would consume 60GB of disk space. Hundreds or thousands of VMs simply compound the storage burden.

NetApp Deduplication: A Hidden Asset for Hyper-V Environments For NetApp customers using server virtualization technologies like Hyper-V, NetApp has reported significant success in reducing customers’ storage capacity requirements -- often by well over 50%. One secret to this savings lies in the company’s deduplication feature for primary storage which first became available with the release of NetApp’s Data ONTAP 7.2.2 operating system for its storage systems. Originally known as A-SIS (for advanced single instance storage) deduplication, the company now refers to this functionality as simply Deduplication for FAS. This feature is available for no added cost on NetApp’s current line of FAS systems.

Building on the Fundamentals of NetApp Snapshot and the WAFL Filesystem.

NetApp deduplication is an integral feature within Data ONTAP and the core NetApp WAFL filesystem. According to one NetApp article,8 NetApp deduplication functionality follows much the same process as existing NetApp Snapshot™ technology. They explain this is one reason deduplication accounts for relatively low system overhead during operation. Specifically, NetApp deduplication builds on the concept of “multiple block referencing” developed years ago with NetApp snapshots. This architecture allows a single WAFL data block to represent multiple point-in-time copies. NetApp’s deduplication then adds the ability to identify unique vs. duplicate “digital fingerprints” associated with each 4KB block of data.

VM 1

VHD file(s)

App.

App. Data

Misc.

OS

VM 2

VHD file(s)

App.

App. Data

Misc.

OS

D D

D D

D D

D D

D D

D D

D D

D D

D D

D D

D D

D D

D D

D U

U D

D D

D U

D D

U D

D D

D D

D D

U U

U U

D U

U U

U D

U U

U U

D U

U U

U U

U U

VM 3

VHD file(s)

App.

App. Data

Misc.

OS

D D

D D

D D

D D

D D

D D

D D

D D

D D

D D

D D

U D

D D

D D

D D

D U

D D

D D

U D

D U

D D

D D

U U

U U

U D

U U

U U

D U

U D

U U

U U

U D

U U

U U

U U

U U

U U

U U

U U

U U

U U

U U

U U

U U

U U

U U

U U

U U

U U

U U

U U

U U

U U

U U

U U

The Storage Writer




NetApp deduplication accesses and compares Data ONTAP metadata already containing the digital fingerprints associated with each 4K data block. After identifying duplicate fingerprints and performing a byte-by-byte scan to ensure the physical blocks are true duplicates, a simple block-redirect process moves the duplicate block to the “free pool”. It also references the duplicate block’s data pointer to the original, previously stored block. The NetApp article refers to this process as “using one ‘physical’ data block to represent many ‘logical’ data blocks.”

How Deduplication Works: An Overview

Since the digital fingerprints for each data block already exist in the WAFL filesystem, it’s an easy process to run NetApp deduplication. Instead of being focused on just backup data sets, NetApp deduplication runs on any data blocks contained within a deduplication-enabled NetApp flexible volume (also known as FlexVol™). NetApp deduplication performs the following operational steps:

1. Lists the digital fingerprints of block data residing in the deduplication-enabled volume. 2. Uses the fingerprints to compare and identify data blocks which are duplicates. According to

NetApp, this keeps system overhead low during the process. 3. After identifying two or more duplicate blocks, it follows up with a byte-by-byte scan to confirm

data in the block is an exact duplicate. (According to NetApp, this added scan removes the risk of “false positives” reported by other deduplication technologies while still minimizing the performance overhead required to perform deduplication.)

4. Moves any reference data pointers from the duplicate blocks to the remaining, “single instance” data block, while subsequently marking the duplicate blocks as “free.”

Operation with Hyper-V

Some application environments lend themselves more readily to higher capacity savings and space reclamation with NetApp deduplication. In general, environments with a lot of duplicate data (like that found with Hyper-V) are often excellent candidates for deduplication, as shown in Figure 2.

Conceptual Example: Storage of File Data for Five Hyper-V Virtual Machines Without NetApp Deduplication With NetApp Deduplication

Source: NetApp

Figure 2. NetApp Deduplication with Hyper-V.

The Storage Writer




Some Important Points About Use with Hyper-V

Unlike many deduplication solutions on the market, NetApp promotes its deduplication functionality as being able to cover everything -- from deduplicating primary volumes to deduplication of secondary backups, archives and replicated data sets. The deduplication process also has a minimal performance impact while it runs, not requiring either the use of complex hash algorithms or look-up tables to preserve the data's “single-instance” status. NetApp deduplication features often touted include:

First-to-market innovation with the ability to deduplicate data on primary storage volumes, as well as data being backed up or remotely replicated to secondary backup or archival volumes.

The ability to deduplicate data residing on both new FlexVol volumes, as well as existing FlexVol volumes. This means flexible volumes with pre-existing applications and data, such as those commonly found in existing Hyper-V environments, can still take advantage of deduplication.

The ability to launch the deduplication process manually, on an automated (off-peak) schedule, or when the data capacity of a volume reaches a certain threshold. Deduplication can also be just as easily disabled.

Hyper-V Caveats: While reports of high space savings with Hyper-V are common among NetApp users, NetApp cautions that a Hyper-V customer's specific deduplication “mileage”, space savings and performance impact may vary. The OS and application portions of a VHD file are likely to reclaim the most space from NetApp deduplication, especially for applications which are near replicas of each other. In contrast, a VM’s unique application data and temporary files will not fare as well during the deduplication process. Certain applications also may fare much better than others.

General Tips: To maximize deduplication performance and space savings in Hyper-V environments, NetApp offers the following tips9, 10: Enable NetApp LUN thin provisioning (via NetApp

FlexVol) for LUNs associated with Hyper-V environments. Also, pay close attention to how you configure the LUNs in order to maximize deduplication space savings. 11

Since you can have files from multiple VMs on the same NetApp volume, group similar operating systems and applications into the same NetApp flexible volumes (FlexVols). o In some cases, it may increase savings by

Efficiency Profile: Nearly 70% Space Savings

-------------------------------------------- One NetApp customer20 achieved excellent

space savings through its use of NetApp deduplication with Hyper-V. This allowed the customer to subsequently defer added storage purchases for two years.

A&A Contract Customs Brokers Ltd. (AACB) is a Canadian/U.S. customs and freight broker that had nearly run out of storage space by the time it tried NetApp deduplication on both its virtualized (roughly 20 virtual machines) and non-virtualized application data.

Dan Morris, AACB’s systems engineer, reported the results: He was able to reclaim close to 70% of the original disk space consumed. “At last check, we saved over 1TB and now use only 400GB to store all our Hyper-V guests,” he said.

AACB’s Microsoft Data Protection Manager backups also showed high savings with NetApp deduplication when used for the DPM storage pool. “For what would usually be 2.5TB of backup data, we now need just 800GB of disk space,” Morris stated. Five hundred gigabytes of SQL Server® data from the company’s SpectorSoft monitoring system was also being stored on NetApp. Deduplication even brought that amount down to 300GB.

The Storage Writer




separating the operating system and application binaries/data onto two separate VHDs. Then store each of these VHD files associated with the same VM on its own separate NetApp FlexVol.

o In addition, for folks deploying large amounts of Microsoft Windows OSes, consider separating the Windows Paging file and temporary data onto its own VHD. Store these in another NetApp FlexVol, separately from the FlexVols containing OS and App content.

Enable NetApp deduplication on the FlexVol volumes containing the VHDs where the OS and Application content resides.

If you have separated the Microsoft Windows OS Paging file and temporary data onto its own VHD and NetApp FlexVol, do not enable NetApp Deduplication on that volume.

Learn how NetApp deduplication works on volumes with NetApp Snapshot data.12

NetApp advises users to first assess the potential impact of running NetApp deduplication on each application used in their Hyper-V environment. Applications requiring sequential reads or writes, for instance, do not make good candidates for deduplication. For details, see the FAQ on NetApp deduplication for FAS13 as well as a related deduplication article14 from NetApp’s Tech OnTap Newsletter.

Lessons in Hyper-V Savings with NetApp Deduplication Organizations that use NetApp deduplication with Hyper-V tend to report significant storage capacity savings when performing key activities. These include:

Creating or cloning virtual machines Reclaiming storage space on existing storage

volumes Performing disaster recovery to another location The following subsections offer more details.

Efficiency Profile: Over 50% Saved on OS VHDs

-------------------------------------------- Avanade is another NetApp customer21 that

has seen promising space savings through the use of NetApp deduplication in their Hyper-V environment.

The Seattle-based enterprise IT consulting firm has over 100 virtual machines spread across its production, test and development environments and had been exploring how best to grow the environment further while still containing the future cost of disk space, power and cooling.

Avanade Senior Systems Engineer Andy Schneider decided to try a test run of NetApp deduplication in Avanade’s development and test environment. After starting the deduplication process one afternoon on a NetApp volume containing Microsoft Windows 2008 child OSes, Schneider went to lunch. He returned to find the process had completed and freed over 50% of the original storage capacity used by the child OS VHD files.

In a follow-up webinar,22 Schneider elaborated further. “One of the beauties of using NetApp is the deduplication feature they’ve introduced. You can go in and basically enable [deduplication] on a volume in NetApp,” he said, noting, “W e’ve seen in some Dev and Test [virtual machine environments] that we’ve been running up to 56% savings. This continues to grow as we add more and more similar virtual machines to that existing volume on our NetApp. So that was a huge benefit for us.”

The Storage Writer




Lesson #1: Creating or Cloning Virtual Machines

Many organizations use Microsoft System Center Virtual Machine Manager (SCVMM) to create a new VM from a prior VM “template” or clone an existing VM with pre-installed operating system and application. Once the new VM creation or cloning process is completed, new VHDs will contain much common content. Running NetApp deduplication on a flexible volume with these VHD files can yield high capacity savings. NetApp estimates typical capacity savings of over 50% using NetApp deduplication in Hyper-V environments. 15 Some customers even report storage capacity savings in excess of 80%.

Lesson #2: Reclaiming Storage Space on Existing Storage Volumes

The process involved in reclaiming storage space “freed” by the NetApp deduplication process is fairly straightforward. In general, NetApp deduplication is designed to examine and discover all duplicate blocks in a NetApp FlexVol volume, then return any duplicates it finds back to the storage system. In effect, these duplicates are returned to the volume as “free storage” blocks for reuse. Any data pointers that previously pointed to the duplicate blocks will now point, instead, to the one remaining, identical block on the volume. Reclaiming Freed Blocks in a LUN Environment. A unique benefit of combining Microsoft virtualization with NetApp storage is the ability to reclaim space on the fly without impacting applications. There is a process involved in viewing and reusing previously freed storage blocks in a LUN environment, especially if you are using thin provisioning on your flexible volumes and your LUNs have been configured with certain space guarantees. NetApp demonstrates much of this process of reclaiming and reusing space within a LUN in its YouTube demo, “Honey, I Shrunk the LUNs,” at http://www.youtube.com/watch?v=ibOtcWrUkPw.16 NetApp’s deployment and implementation guide also describes how NetApp deduplication and space reclamation work under five different LUN configurations.17

Efficiency Profile: From 9PB to 0.65TB

-------------------------------------------- The Hands-On Lab (HOL) at Tech Ed Taiwan

2008 faced a BIG storage problem: Being able to offer each of the conference’s 2600+ participants what amounted to 3 TB of online courseware and curriculum running virtually on Hyper-V. Project organizers wanted students to be able to log in to any computer in the hands-on lab and access their own 3TB course data from their own virtual disk. The virtual disk would just appear to them as a local disk.

Working out the numbers, organizers calculated the total storage needed to accommodate this scenario with so many users. It came to a staggering 9PB, or 9,000 TB!

According to Clive Chiang, a NetApp consulting system engineer involved in the project, such a high capacity need would have been impossible to accommodate with the space available in the conference center facility. By his estimates, it would have required roughly 231 storage cabinets which would have amounted to more space than the entire first floor of the Taipei International Conference Center.

Instead, using primarily NetApp deduplication functionality and NetApp thin provisioning, they were able to drop the original 9,000 TB capacity requirement down to just 0.65 TB of storage space for all 2600+ users. According to Chiang, this amounted to a space savings of up to 99.9928%.23

http://www.youtube.com/watch?v=ibOtcWrUkPw

The Storage Writer




Lesson #3: Performing Disaster Recovery to Another Location

As the number of production applications running in virtual machines increases, the need to replicate the underlying VM data to another location becomes more important as organizations seek to protect their most critical systems from a local or site-wide disaster. Without deduplication, however, off-site replication may turn out to be a costly affair because of the large amounts of duplicate data needing to be transmitted “across the wire” to the remote site. In an article by NetApp Product and Partner Engineer Jeremy Merrill,18 Merrill described the following deduplication scenario for virtual server environments using two NetApp systems -- one local and one remote. This scenario involves the use of NetApp SnapMirror for remote replication between the sites. The next three basic steps (and accompanying figure) depict how the same process also works for Hyper-V environments: 1. The user deduplicates primary Hyper-V volumes at the main data center 2. Once Hyper-V data is deduplicated, SnapMirror transmits the deduplicated 4K blocks of data (or

subsequent changes to the original blocks) to the remote site. Each 4K block transmitted is unique. Duplicates are not transmitted, significantly shrinking the overall transmission costs.

3. Once the Hyper-V replicated data resides at the remote site, you do not need to run deduplication again since all the benefits of deduplication are “inherited” from the primary system.

Source: Adapted from earlier NetApp diagram.

Figure 3. Deduplication Savings are “Inherited” by the Target Remote System During SnapMirror Replication.

“Inherited” Deduplication Over 50% Space Savings

Remote Data Center

Hyper-V Hosts

SAN/ NAS

Primary Data Center

“Golden” VM Templates +

Virtual Machine Clones

NetApp FAS System

NetApp FAS System

SnapMirror Replication

NetApp Deduplication Over 50% Space Savings

The Storage Writer




NetApp Deduplication: How to Begin New NetApp customers can automatically take advantage of NetApp deduplication functionality. Existing NetApp customers wanting to explore use of NetApp deduplication in their physical or virtual environments just need to add two “no-charge” software licenses to their NetApp system: A NetApp NearStore license and a deduplication license. NetApp V-Series systems also offer use of NetApp deduplication functionality for data stored on other vendor storage arrays -- such as many produced by Hitachi Data Systems, EMC, HP and others. Using a simple CLI command, you can enable deduplication on any NetApp flexible volume. Other commands allow you to specify when and how often you want to run deduplication on the volume. Once deduplication is enabled on each volume, the volume’s digital fingerprints will begin to be gathered and made ready for the deduplication process. To learn more, go to: http://www.netapp.com/us/products/ and select the “Deduplication” link under the category of “Platform OS/Features”. You can also refer to the End Note references at the end of this brief or to NetApp Technical Report 3505, “NetApp Deduplication for FAS Deployment and Implementation Guide.”19

About This Report

This brief was commissioned by NetApp and involved the author conducting interviews with NetApp personnel and NetApp customers. Research was also performed on the part of the author via NetApp-internal sources and publicly available sources both within and outside of NetApp. Where possible, efforts were made by the author to attribute sources referenced throughout the brief and verify the accuracy of information presented herein. The author, Michele Hope, of the Storage Writer, Inc. is an independent storage industry commentator, writer and reporter with over 20 years of experience writing about IT topics. Michele’s storage reports have appeared in venues ranging from InfoStor and Network World to FedTech Magazine and SearchStorage.com. She can be reached at [email protected].

http://www.netapp.

The Storage Writer


NOTICE: The information, advice and/or recommendations made by the The Storage Writer, Inc. may include personal opinions of The Storage Writer, Inc. and others, which we believe to be accurate and reliable. Such information, advice and recommendations are made without warranty of any kind. The Storage W riter, Inc. assumes no responsibility or liability for any damages whatsoever (including incidental, consequential or otherwise), caused by your use of, or reliance upon, the information and recommendations presented herein, nor for any inadvertent errors which may appear in this document. All product names used and mentioned herein are the trademarks or registered trademarks of their respective owners. Copyright © The Storage Writer, Inc. 2009. All Rights Reserved. Page 13

The Storage Writer, Inc. Storage Insights, Technical Research and Reports Phone: (770) 643-2050 E-mail: [email protected] Web: www.thestoragewriter.com

End Notes

1 The NetApp Virtualization Guarantee program guarantees those using NetApp storage in their virtual server environments will use 50% less storage space than traditional storage. (NetApp’s expected deduplication savings is a key component behind the NetApp Virtualization Guarantee program.) Program details are available at http://www.netapp.com/us/solutions/infrastructure/virtualization/guarantee.html 2 “NetApp customers implement deduplication technology on more than 37,000 systems worldwide,” NetApp press release, May

13, 2009, at http://www.netapp.com/uk/company/news/news-rel-20090513-uk.html. 3 “Forrester Survey Exposes Storage Virtualization Truths,” by Amy Newman, ServerWatch, March 4, 2009,

http://www.serverwatch.com/news/article.php/3808541. 4 “Struggling with Super-Sized Storage,” by John Edwards, IT World, April 6, 2009,

http://www.itworld.com/storage/65914/struggling-supersized-storage. 5 "Opinion: De-duplicating VM ware," by Jim Damoulakis, Computerworld, March 12, 2008,

http://www.computerworld.com/action/article.do?command=viewArticleBasic&articleId=9068059. 6 "Creating virtual machines in VMM," Microsoft System Center Virtual Machine Manager TechCenter, Microsoft Corporation,

Oct. 21, 2008, http://technet.microsoft.com/en-us/library/cc764227.aspx. 7 “Windows Server 2008 R2 & Hyper-V Server 2008 R2 RTM!!!!,” Blog entry, July 22, 2009, Microsoft Windows Virtualization

Team Blog, http://blogs.technet.com/virtualization/archive/2009/07/22/windows-server-2008-r2-hyper-v-server-2008-r2-rtm.aspx.

8 "The Perfect Virtual Marriage: Deduplication and VMware," by Larry Freeman and Bill May, NetApp Tech OnTap newsletter, January 2008, http://partners.netapp.com/go/techontap/matl/0108techontap.pdf.

9 Updated every few months by NetApp, several of these guidelines are also found in “NetApp and Microsoft Virtualization Best Practices,” NetApp Technical Report 3702, September 2008, NetApp Corporation, http://media.netapp.com/documents/tr-3702.pdf. January 13, 2009 blog posting by NetApp virtualization blog

10 "Savings extended for Hyper-V and XenServer," blog posting by Chaffie McKenna, January 13, 2009, at NetApp’s “The Virtualization Effect” blog, http://blogs.netapp.com/virtualization/2009/01/savings-extende.html.

11 "NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide," NetApp Technical Report 3505, Section 4.15, by Carlos Alvarez, April 16, 2008, NetApp Corporation, http://media.netapp.com/documents/tr-3505.pdf.

12 Ibid. Refer to Section 4.1. 13 "NetApp Deduplication for FAS Frequently Asked Questions," NetApp Corp., March 2009,

http://communities.netapp.com/docs/DOC-1060. 14 "Deduplication Comes of Age, Version 2," by Blake Lewis, NetApp Tech OnTap Newsletter, July 2008,

http://www.netapp.com/us/communities/tech-ontap/dedupe-0708.html. 15 See Reference Note #1. 16 “Honey, I Shrunk the LUNs... (a Hyper-V on NetApp Demo),” NetAppTV, September 11, 2009, YouTube,

http://www.youtube.com/watch?v=ibOtcWrUkPw. 17 See Reference Note #11. 18 “Combining Deduplication and VMware DR,” by Jeremy Merrill, NetApp Tech OnTap Newsletter, January 2008,

http://partners.netapp.com/go/techontap/matl/dedup_VM_DR.html. 19 See Reference Note #11. 20 Customer profile excerpted and summarized from NetApp Success Story on AACB, which is available at:

http://media.netapp.com/documents/aacb.pdf. 21 Customer profile excerpted and summarized from NetApp Success Story on Avanade, which is available at

http://media.netapp.com/documents/avanade.pdf. 22 Excerpted from “Benefits,” Slide 17, within “NetApp Storage Solutions for Microsoft Hyper-V - Optimize Your Virtual

Infrastructure Using 50% Less Storage,” a NetApp OnDemand Webinar available at: http://www-download.netapp.com/edm/TT/WOD/WOD20081002/index.html.

23 “NetApp successfully helps Microsoft reduce overall storage capacity up to 99.9928% for Hands-on Lab of Tech Ed 2008 Taiwan,” NetApp Press Release, Oct. 20, 2008.

http://www.netapp.com/us/solutions/infrastructure/virtualization/guarantee.html

http://www.netapp.com/uk/company/news/news

http://www.serverwatch.com/news/article.php/3808541

http://www.itworld.com/storage/65914/struggling

http://www.computerworld.com/action/article.do?command=viewArticleBasic&articleId=9068059

http://technet.microsoft.com/en

http://blogs.technet.com/virtualization/archive/2009/07/22/windows

http://partners.netapp.com/go/techontap/matl/0108techontap.pdf

http://media.netapp.com/documents/tr

http://blogs.netapp.com/virtualization/2009/01/savings

http://media.netapp.com/documen

http://communities.netapp.com/docs/DOC

http://www.netapp.com/us/communities/tech

http://www.youtube.com/watch?v=ibOtcWrUkPw

http://partners.netapp.com/go/tec

http://media.netapp.com/documents/aac

http://media.netapp.com/documents/avanade.pdf

Documents

Secrets to Shrinking Your Storage for Microsoft Windows ...download.microsoft.com/documents/uk/enterprise/85... · company·s deduplication functionality has been implemented in over