lab validation report -...

Lab Validation Report Hewlett Packard StoreEasy 5000 Storage

Robust, Simple, High Performance Midrange file storage

By Tony Palmer, Senior Lab Analyst and Mike Leone, Lab Analyst

December 2012 © 2012, Enterprise Strategy Group, Inc. All Rights Reserved.

Lab Validation: HP StoreEasy 5000 Storage 2

© 2012, Enterprise Strategy Group, Inc. All Rights Reserved.

Contents

Introduction .................................................................................................................................................. 3 Background ............................................................................................................................................................... 3 HP StoreEasy 5000 Storage ...................................................................................................................................... 4 HP StoreEasy 5000 Storage Software Features ........................................................................................................ 6

ESG Lab Validation ........................................................................................................................................ 8 Getting Started with HP StoreEasy 5000 .................................................................................................................. 8 HP StoreEasy 5000 Performance ............................................................................................................................ 13 HP StoreEasy 5000 Efficiency ................................................................................................................................. 16

ESG Lab Validation Highlights ..................................................................................................................... 21

Issues to Consider ....................................................................................................................................... 21

The Bigger Truth ......................................................................................................................................... 22

Appendix ..................................................................................................................................................... 23

All trademark names are property of their respective companies. Information contained in this publication has been obtained by sources The Enterprise Strategy Group (ESG) considers to be reliable but is not warranted by ESG. This publication may contain opinions of ESG, which are subject to change from time to time. This publication is copyrighted by The Enterprise Strategy Group, Inc. Any reproduction or redistribution of this publication, in whole or in part, whether in hard-copy format, electronically, or otherwise to persons not authorized to receive it, without the express consent of the Enterprise Strategy Group, Inc., is in violation of U.S. Copyright law and will be subject to an action for civil damages and, if applicable, criminal prosecution. Should you have any questions, please contact ESG Client Relations at 508.482.0188.

ESG Lab Reports

The goal of ESG Lab reports is to educate IT professionals about data center technology products for companies of all types and sizes. ESG Lab reports are not meant to replace the evaluation process that should be conducted before making purchasing decisions, but rather to provide insight into these emerging technologies. Our objective is to go over some of the more valuable feature/functions of products, show how they can be used to solve real customer problems and identify any areas needing improvement. ESG Lab's expert third-party perspective is based on our own hands-on testing as well as on interviews with customers who use these products in production environments. This ESG Lab report was sponsored by Hewlett-Packard.



Introduction

This ESG Lab Validation documents the results of ESG Lab hands-on testing of the HP StoreEasy 5000 Storage file appliance with an emphasis on the simplicity, performance, and synergies with Microsoft Windows Server environments. Testing was designed to validate the StoreEasy 5000 as a robust, viable midrange file storage appliance. Of special interest was high availability, SMB 3.0 support, “cluster-in-a-box” features as well as the “green” benefits delivered in StoreEasy 5000’s compact 3U form factor.

Background

IT managers within organizations of all sizes are challenged by the cost and complexity associated with managing growing volumes of digital data. These challenges are particularly vexing for IT managers within midmarket organizations (100-999 employees) due to a shortage of manpower and dedicated storage expertise compared to their peers within enterprise-class organizations (1,000 employees or more). More than half of IT managers within midmarket companies surveyed by ESG (60%) reported managing 25TB or more of data1 and they’re struggling to keep up with data growth. Nearly half of the respondents to a separate ESG research survey (46%) reported that data is growing at an annual rate of 20% or more.2 For those organizations, data volume is doubling every two to five years.

Faced with these challenges, it’s no surprise that the top storage challenges reported by IT managers would include rapid growth and management of unstructured data, hardware costs (CAPEX), and poor performance, as shown in Figure 1.3 It should also be noted that the need to support growing virtual server environments and the management, automation, and placement of data made the list of top ten challenges as well.

Figure 1. Top Ten Storage Environment Challenges

Source: Enterprise Strategy Group, 2012.

1 Source: ESG Research Brief, Disk-based Storage Capacity Trends, September 2012.

2 Source: ESG Research Report, Trends in Data Protection Modernization, August 2012.

3 Source: ESG Research Report, 2012 Storage Market Survey, due to be published October 2012.

15%

15%

17%

17%

17%

19%

19%

20%

25%

25%

25%

39%

39%

40%

0% 10% 20% 30% 40% 50%

Poor performance (I/Os)

Poor performance (throughput)

File system expansion

Device management

Discovery, analysis and reporting of storage usage

Lack of skilled staff resources

Management, optimization & automation of data placement

Staff costs

Data migration

Need to support growing virtual server environments

Running out of physical space

Data protection

Hardware costs

Rapid growth and management of unstructured data

In general, what would you say are your organization’s biggest challenges in terms of its storage environment? (Percent of respondents, N=418)



HP StoreEasy 5000 Storage

HP has designed the StoreEasy 5000 as a highly available, high performance, turnkey file and application storage appliance. The StoreEasy 5000 is a two-node clustered controller based on HP Gen8 blade server hardware with high density storage. Windows Storage Server 2012, with SMB 3.0 enhances the reliability, availability, manageability, and performance of StoreEasy 5000 and enables customers to deploy storage for virtualized application workloads such as Exchange and Microsoft SQL Server. Encryption and signing are included to protect data in motion while data at rest is protected by BitLocker drive encryption. The StoreEasy 5000 leverages Windows Storage Server 2012’s radically improved corruption correction and file system self-healing. HP has addressed uptime concerns with continuous health monitoring, advanced memory protection, hot plug redundant fans and power, as well as a three-year onsite warranty. Figure 2 depicts the StoreEasy family, from entry-level appliances through enterprise ready gateways.

Figure 2. HP StoreEasy Storage Family

HP StoreEasy storage is designed to enable organizations to focus on the business:

Converged—HP StoreEasy Storage supports converged block and file protocols for secure and highly available file consolidation in an easy to manage package.

Efficient—With new deduplication features combined with support for hundreds to thousands of concurrent users and multiple workloads, HP StoreEasy Storage is cost-efficient and easy to manage.

Secure—Hp StoreEasy Storage offers built-in encryption, sophisticated access controls and the ability to run antivirus and backup on the system to keep organizations' data secure whether it’s at rest, in flight or mobile.

Highly Available—HP StoreEasy Storage provides active-active clustered controllers, transparent failover, and non-disruptive upgrades to ensure continuous availability of data to users, physical and virtualized servers, and applications.

Built on HP ProLiant Gen8 hardware with advanced storage features, the HP StoreEasy 5000 is designed as a high availability active-active dual controller storage appliance with up to 48TB in just 3U of rack space that is expandable to hundreds of terabytes with add-on SAS disk shelves. Windows Storage Server 2012, x64 Standard Edition is pre-installed on all HP StoreEasy Storage models(1000, 3000 and 5000) and StoreEasy file storage systems



provide highly available cluster architecture with redundant components and transparent failover. Figure 3 illustrates the architecture. The HP StoreEasy 5000 storage controllers are built from HP ProLiant BL460c G7 blade servers, and support for up to five SAS expansion chassis enables the system to scale in both capacity and performance. HP StoreEasy Storage with WSS2012 is designed for a flexible mix of configurations and includes advanced data management tools.

Figure 3. HP StoreEasy 5000 Storage Architecture

The architecture of the system only tells part of the story. HP StoreEasy Storage combined with WSS2012 enjoys a number of advantages over non-windows file storage in a distributed, heterogeneous IT environment.

Distributed File System Namespace (DFS-N) Enables organizations to group shared folders that are located on different servers into logically structured namespaces. Each namespace appears to users as a single shared folder with a series of subfolders even though the namespace can consist of numerous file shares that are located on different servers and in multiple sites. This enables HP StoreEasy 5000 to virtualize file storage environments and present a single share to users, no matter where the data resides. Proprietary file storage devices that run non-windows operating systems can participate in DFS-N as 'leaf' nodes, or content servers only.

Distributed File System Replication (DFS-R): Enables efficient replication of folders across multiple servers and sites. DFS Replication uses the remote differential compression (RDC) algorithm. RDC detects changes to the data in a file, and it enables DFS Replication to replicate only the changed file blocks instead of the entire file. Proprietary file storage devices that run non-windows operating systems must use their own replication technologies and cannot participate in DFS-R.

BranchCache is designed to decrease wide area network traffic between computers in branch offices and centrally located servers. BranchCache applies deduplication and optimization to data transferred over the WAN to a branch office to offer faster file download times and reduced bandwidth consumption for remote users. Proprietary file storage devices that run non-windows operating systems may participate in branch cache in a limited role (Content Server, for example), HP StoreEasy 5000 Supports both BranchCache Content Server as well as BranchCache Hosted cache mode natively.

HP StoreEasy 5000 is built on proven, standard HP hardware. HP ProLiant Blade servers, HP qualified disk drives, and rock solid HP enclosures provide a cost effective, robust platform to deploy a scalable file storage solution without proprietary hardware. This enables organizations to use a common spares pool across their infrastructure, enabling economies of scale and quicker response to issues.



HP StoreEasy 5000 delivers performance and reliability while offering built-in support for both file (SMB and NFS) and block (iSCSI) protocols, which allows for a flexible mix of configurations unlike many proprietary file storage systems which often charge license fees for each protocol.

HP StoreEasy 5000 enables organizations to leverage existing expertise in Windows Server management tools including Active Directory, PowerShell and System Center.

HP StoreEasy Storage Software Features

HP StoreEasy Storage provides powerful technologies that effectively address converged storage requirements with agility and efficiency. New and improved features provide high performance and scalability that create a dynamic infrastructure which scales predictably to satisfy even the most challenging workloads, while offering the flexibility to adapt to ever-changing business needs. New enterprise-class storage features and functions are being delivered on cost-effective, industry standard hardware:

New features for all models in the StoreEasy family:

Clustering–is now included in the WSS2012 Standard Edition–With no Enterprise Edition for WSS 2012, licensing is simplified

New Protocols–SMB 3.0, NFS v4.1–Provide improved performance, application support, and availability.

Application Shared Storage–over SMB 3.0 for Hyper-V and SQL Server Support for VMware and NFSv4.1–Enabling VMware virtual machines and other third-party

platforms to connect using NFS while benefiting from StoreEasy continuous availability.

In-box NIC teaming–Provides simple, affordable traffic reliability and load balancing enhancing both availability and performance.

Chunk level Deduplication– Enables StoreEasy Storage to achieve storage efficiency that is significantly higher than was possible using WSS2003 or WSS2008. Previous versions used Single Instance Storage (SIS) to eliminate redundant data. Data deduplication operates at the sub-file level and integrates compression to store more data in less physical space and deliver much higher optimization ratios than were possible with SIS, up to 20:1 depending on data type.

ChkDsk Enhancements–Increased performance for faster scans, with online self-healing capabilities.

New Features for Clustered Systems (StoreEasy 3000/5000)

SMB 3.0 Transparent Failover–

SMB 3.0 Multi-Channel–Facilitates network bandwidth aggregation and fault tolerance for multiple paths between SMB clients and file servers.

Cluster Shared Volume V2 (CSV2)–Shared file storage solution for consolidated storage and application clusters in non-Hyper-V environments. Storage space integration with failover clustering also provides continuously available service deployments.

Cluster Aware Updating (CAU)–Enables seamless updates to StoreEasy Storage clustered controllers. CAU both simplifies the process of patching clustered hosts, and helps minimize downtime by automatically moving services across nodes before applying updates.

New Features for StoreEasy 3000 Gateways (with attached SAN array support)

Thin Provisioning/Trim–Allows administrators to create a larger volume with fewer physical disk resources and enables adding or removing physical disk resources without interrupting the storage service, alerts users to the storage volume usage through the threshold setting, uses a shared storage pool configuration for allocating resources, and increases or decreases the size of the storage pool according to the demand and usage of storage space. Trim capability complements thin provisioning by enabling reclamation of provisioned storage that is no longer needed.



Offload Data Transfer (ODX)–Leverages storage array movement technology, offloading these duties from the StoreEasy 3000, enabling fast and easy migration of large data sets or entire VMs between storage devices without impacting StoreEasy or network activities.

Key Differentiating Features of HP StoreEasy Storage

File Classification Infrastructure– File Classification Infrastructure is a feature introduced in WSS2008 R2. FCI is extensible, customizable, and can support many enterprise content management scenarios, without the need for third party software or content management expertise. File Classification Infrastructure provides insight into data by automating classification processes. Organizations can classify files and apply policies based on their classification. These policies provide powerful tools including dynamic access controls for restricting access to files, file encryption, and file expiration and deletion. Files can be classified automatically using file classification rules or manually by modifying the properties of a selected file or folder. IT administrators can leverage the File Classification Infrastructure to automatically classify files, run reports, and apply policy based on file classification properties. Developers can leverage the FCI API's to create custom file classification extensions

Hosted BranchCache–With WSS2008 R2, BranchCache was introduced to help organizations reduce WAN bandwidth usage by client computers that accessed content in remote main offices. With that release, as with BranchCache in WSS2012, administrators can deploy BranchCache in two modes - distributed cache mode and hosted cache mode. With distributed cache mode, client computers retrieve content from the main office content servers and then cache the content to share it with other clients in the same branch office. With hosted cache mode, a server in the branch office–called a hosted cache server - is used as a central cache for the branch. After retrieving content from the main office content server, client computers then store the content on the hosted cache server, which in turn shares the content with other client computers in the branch office. For HP StoreEasy Storage now based on WSS2012, deploying hosted cache mode is significantly easier, thanks to improvements and new features provided by the BranchCache team for this release.

Anti-Virus–HP StoreEasy Storage provides the ability to run any Windows compatible anti-virus software directly on the platform to protect an organization's file data from viruses and malware. With proprietary file storage devices that run non-windows operating systems, organizations need to purchase additional server hardware to provide this vital service. Some non-windows solutions only support anti-virus scanning on Windows files, potentially leaving UNIX- and Linux-based files vulnerable.



ESG Lab Validation

ESG Lab performed hands-on evaluation and testing of HP StoreEasy 5000 at an ESG facility in San Mateo, California. Testing was designed to demonstrate the simplicity of ordering, deploying and managing the HP StoreEasy 5000 appliance while performance was analyzed using industry standard tools and methodologies. Also of interest were the enhancements to Windows Storage Server 2012 and how smoothly it was integrated with the StoreEasy 5000.

Getting Started with HP StoreEasy 5000

ESG Lab tested the StoreEasy 5000 with a goal of showing the ease of configuration, performance scalability, and cost-effectiveness of the appliance. Figure 4 shows the hardware configuration used during testing. ESG Lab used two HP ProLiant ML 150 G6 server as the host machines running VMware vSphere 5.1. A total of 12 guest VMs running Windows Server 2012 were each configured with two vCPUs and 12GB of RAM each. The host machine was connected via 10Gb Ethernet to an HP StoreEasy 5000 Storage system with 36 900GB small form factor (SFF) 10K SAS drives.

Figure 4. The ESG Lab Test Bed

ESG Lab Testing

ESG Lab followed the HP Quick Start guide to install and configure access to the StoreEasy 5000. The Quick Start guides the user through powering up the chassis, making network connections, and connecting to the HP Integrated Lights-Out console, shown in Figure 5. The iLO console enables administrators to monitor and manage the StoreEasy 5000 enclosures over a standard web browser, and connect to the StoreEasy 5000 node consoles via built-in remote desktop functionality. Once connectivity was established, ESG Lab logged on to the primary node using the iLO remote desktop to complete configuration.



Figure 5. The HP Integrated Lights-Out Console

Next, ESG Lab stepped through the Initial Configuration Tasks, or ICT. The ICT guides the user through configuring the cluster and joining it to an Active Directory domain, as shown in Figure 6.

Figure 6. Creating the Cluster and Joining it to the Domain

All aspects of cluster configuration are handled by the wizard, including creating a witness disk and validating the failover cluster configuration.



Once the Cluster configuration and validation were complete, the easy to follow Create File Server wizard was used to provision and share storage, as highlighted in Figure 7.

Figure 7. Creating a File Server

The default protocol for shares is SMB 3.0. To create an NFS share for use by VMware, ESG Lab selected the Server Manager in the taskbar, browsed to the Shares section, and created a new share. The wizard asked for information regarding server and path, the desired name of the share, authentication method (StoreEasy supports Kerberos v5 with integrity and/or privacy), and permissions. After selecting everything, a confirmation screen is displayed before the volume is created, as shown in Figure 8.



Figure 8. Creating an NFS Export

Finally, ESG Lab mounted the NFS share using the vSphere client Add Storage wizard and began creating virtual machines, seen in Figure 9.

Figure 9. Adding storage to vSphere

Finally, ESG Lab used vMotion to migrate an existing Windows Server 2012 virtual machine into the new datastore, powered it on, and mapped a share from the StoreEasy 5000 as seen in Figure 10.



Figure 10. Running Virtual machines from a StoreEasy-hosted NFS Datastore

All told, ESG Lab had the StoreEasy 5000 cluster up and running and joined to the domain with more than 25TB of storage configured and shared out via NFS for VMware and SMB/CIFS for Windows clients, including configuring the clients to access the shares and cloning the first virtual machine into the NFS datastore in less than 90 minutes.

Why This Matters

ESG research indicates that server virtualization is the top IT priority over the next 12 to 18 months, and more than half of IT managers within midmarket companies surveyed by ESG (60%) reported managing 25TB or more of data4. Nearly half of the respondents to a separate ESG research survey (46%) reported that data is growing at an annual rate of 20% or more, which equates to data volume is doubling every two to five years.5

Less than 90 minutes after ESG Lab started testing, virtualized servers were being hosted on an NFS-shared datastore from a highly available HP StoreEasy 5000 cluster. The solution installed quickly and easily using easy to follow wizards and standard Windows Server tools familiar to IT generalists, requiring no specialized storage knowledge. More than 25TB of usable storage was provisioned and shared out via NFS for VMware and SMB/CIFS for Windows clients quickly and easily.

4 Source: ESG Research Brief, Disk-based Storage Capacity Trends, September 2012.

5 Source: ESG Research Report, Trends in Data Protection Modernization, August 2012.



HP StoreEasy 5000 Performance

ESG Lab tested the improved SMB 3.0 protocol with a goal of showing the ease of configuration, performance scalability, efficiency, and cost-effectiveness in a business environment. The same hardware configuration shown in Figure 4 was used during testing. ESG Lab used HP ProLiant ML150 G6 servers as the host machines running VMware vSphere 5.1. The host machines were connected to the HP StoreEasy 5000 over a 10GbE LAN. Eight guest VMs running Windows Server 2012 were each configured with two vCPUs and 12GB of RAM. The virtual machines mounted shares from the StoreEasy 5000 server through the SMB 3.0 protocol.

ESG Lab Testing

Configuring SMB shares was performed using the same user-friendly wizard used to configure NFS shares. Figure 11 shows the SMB share setup process. By selecting New Share from the Volumes section of Server Manager, the wizard guided ESG Lab through the setup. For this particular test case, ESG Lab selected an SMB share offered specifically for a database application workload. After providing information in regards to the share location, name, and access permissions, a new share was created and presented to the main test server.

Figure 11. Configuring SMB Shares

ESG Lab used an OLTP online brokerage application to simulate the activity of thousands of Microsoft SQL Server 2012 users with a goal of demonstrating the performance, scalability, and efficiency of the StoreEasy 5000 running the SMB protocol. A 1,000 customer database was configured within each of the eight SQL Server virtual machines with a goal of achieving linear scalability for the number of transactions per second as the number of consolidated SQL Server VMs increased from one to eight. The transactions per second and average response time were monitored as the number of customers and VMs increased. The results are summarized in Figure 12 and Table 1.



Figure 12. SQL Server Workload Scalability with the SMB 3.0 Protocol

Table 1. SQL Server Workload Scalability with the SMB 3.0 Protocol

Virtual Machines

Database Scale (Total Customers)

Transactions/Sec (Total)

Average Transaction Response (seconds)

1 1,000 552 0.01

2 2,000 1,072 0.01

4 4,000 2,141 0.02

6 6,000 3,029 0.02

8 8,000 3,863 0.06

What the Numbers Mean

Each of the eight virtual machines was populated with a SQL Server database supporting 1,000 brokerage customers.

The virtual machines were configured with two virtual CPU cores and 12GB of RAM with a goal of balancing the utilization of all the available physical server resources during the peak eight-VM test.

Transactions per second increased from 552 to 3,863 as the number of virtual machines running on a single physical server increased from one to eight.

ESG Lab recorded manageably low average transaction response times between 0.01 and 0.06 seconds, avoiding end-user-perceived slow response times for transactions taking longer than three seconds.

Network utilization remained remarkably low, with utilization reaching a maximum of 3.5%

ESG Lab also tested performance of the NFS protocol by running applications on virtual machines with disks provisioned from the NFS datastore shared out by the StoreEasy 5000. ESG used the Iometer workload generation utility to simulate workloads generated by typical business applications and the same OLTP online brokerage application used in the previous test. 6 Each workload was run in one Windows Server 2012 virtual machine with 2 vCPUs , 12GB of RAM, and a 30GB volume provisioned from the NFS datastore. The results are summarized in Figure 13 and Table 2.

6 Configuration details are detailed in the appendix.

0.0

0.2

0.4

0.6

0.8

1.0

0

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

1 2 4 6 8 Ave

rage

Tra

nsa

ctio

n R

esp

on

se (

sec)

Tran

sact

ion

s/se

c

Number of Hyper-V VMs

SMB 3.0 SQL Server Workload Scalability (OLTP Workload, Windows Server 2012, SQL Server 2012)



Figure 13. Mixed Workload Scalability with the NFS 4.1 Protocol

Table 2. SQL Server Workload Scalability with the SMB 3.0 Protocol

Virtual Machines Database Scale

(Total Customers)

Exchange 2010 Scale

(Total Mailboxes)

Media Server Scale

(Total MBps)

Average Response Time

(ms) 3 1,000 4,855 110 11.2

6 2,000 8,320 170 12.2

9 3,000 10,470 264 17.1

12 4,000 12,790 359 17.4


Each of the 4 SQL Server virtual machines was populated with a SQL Server database supporting 1,000 brokerage customers.

The virtual machines were configured with a 30GB data volume provisioned from the NFS mounted datastore.

All workloads scaled nearly linearly as the number of virtual machines running on two physical servers was increased from three to 12.

ESG Lab recorded consistently low average disk response times between 11.2 and 17.4 milliseconds. OLTP transaction response times stayed below one tenth of a second, again avoiding end-user-perceived

slow response times for transactions taking longer than three seconds. Exchange response times ranged from 7.9 to 17.5 ms, below the Microsoft-recommended threshold for

Exchange database latency of 20 milliseconds.

3 6 9 12

Number of Virtual Machines

Mixed Workload Performance

OLTP 4K Exchange 2010 Media Server



Why This Matters

With less tolerance for poor application performance, it’s no surprise that many organizations have resisted moving business-critical applications to virtual servers for fear that workload aggregation will slow performance. Traditionally, remote storage options for server virtualization platforms were limited to costly, hard to provision SAN solutions for guests or less expensive storage options without advanced storage functionality. The enhanced performance and efficiency of the SMB 3.0 protocol in the HP StoreEasy family allows organizations to save money while getting great performance on industry standard hardware.

ESG Lab validated that the improvements to the SMB protocol allow organizations to deploy applications like SQL Server 2012 without a need for concern. Deploying an extremely cost-effective solution that leverages the new benefits of the SMB 3.0 protocol, consolidation through VMware vSphere, and easy provisioning of the HP StoreEasy 5000, more than adequate performance was achieved from an enterprise SQL Server 2012 application workload. The number of transactions/sec scaled nearly linearly from 552 to 3,863 while response times remained low as the number of VMs scaled from one to eight.

In a VMware vSphere environment, the HP StoreEasy 5000 showed excellent NFS performance as well, scaling nearly linearly with mixed business -critical workloads on 12 virtual machines using storage provisioned from an NFS-mounted datastore.

Finally, very little bandwidth was consumed, giving more freedom for network configurations within IT infrastructures. With the enhanced benefits of SMB 3.0 and NFS 4.1, organizations can achieve similar performance and high availability of many enterprise-class storage deployments.

HP StoreEasy 5000 Efficiency

The HP StoreEasy 5000 now includes data deduplication as a standard feature. This new feature is a highly scalable and efficient solution that allows businesses to store more data in less physical space. With the added capabilities of capacity optimization, non-intrusive performance scalability, and reliability from the perspective of data integrity, StoreEasy allows users to store and access data as efficiently as possible. Deduplication is included with the StoreEasy family at no extra cost, and is scalable and non-intrusive. Using a combination of chunking and compression technologies, StoreEasy Storage applies state-of-the-art deduplication to each volume, freeing up storage capacity.

ESG Lab Testing

ESG Lab tested the new StoreEasy deduplication capabilities with a focus on capacity savings and potential deduplication overhead. Two different data sets were tested:

File Server / Home Directory – The volume size was 900GB and consisted of mostly small files such as Microsoft Office documents and PDFs. This data set is considered to be moderately deduplicateable due to the size and variability of file content.

Operating System Images – The volume size was 3TB and consisted of mostly large files such as pre-configured VHDs containing Server 2008, Server 2012, and SQL Server 2012. This data set is considered to be highly deduplicateable due to the combination of size and the amount of common data.

ESG Lab’s first objective was to understand the ease of configuring deduplication on a particular volume. Setting up deduplication was made simple through a single window that was accessed from Server Manager. The short process is shown in Figure 14. ESG Lab selected the volume in which to deduplicate and accessed the Deduplication Settings window. The window allowed for either a quick and easy setup, or offered the ability to create a deduplication schedule for more complex storage environments.



Figure 14. Configuring Deduplication

Next, ESG Lab used Powershell commands to initiate deduplication jobs on the two volumes specified above. 7 The command was issued to one volume at a time. There are two deduplication modes: throughput mode and background mode. Throughput mode is meant to consume more server resources (CPU and memory) and complete faster. Background mode is meant to consume less server resources and complete slower. With a goal of showing capacity savings, ESG Lab used throughput mode to quickly deduplicate each volume. The capacity savings are shown in Figure 15.

Figure 15. Server 2012 Deduplication Capacity Savings

The File Server volume decreased in consumed capacity from 900GB to 648GB, yielding a savings of 28%. HP guidelines suggest a typical user could see between 25-60% capacity savings with similar datasets. The volume consisting of OS VHDs decreased in consumed capacity from 3TB to 57GB, yielding a savings of 98%. ESG Lab viewed the deduplication rate through Server Manager, as shown in Figure 16.

7 Listed in appendix.

0

500

1000

1500

2000

2500

3000

3500

File Server Data OS VHDs

Co

nsu

me

d C

apac

ity

(GB

s)

StoreEasy Deduplication Capacity Savings

Before Deduplication After Deduplication



Figure 16. Deduplication Rate through Server Manager

Impressed with the capacity savings, ESG Lab next tested the overhead of rehydrating files in two different ways: opening small files and copying large files. Using a Powershell script, ESG Lab selected four different file counts consisting of Microsoft Office files. Each test opened the selected number of files from a volume where the files were not deduplicated and then from a volume where they were deduplicated. Processor utilization and elapsed time were monitored and the results are shown in Figure 17 and Table 3.

Figure 17. Deduplication Overhead – Opening Multiple, Small Files

0 5 10 15 20 25 30 35 40

125

83

22

1

Elapsed Time to Open File(s) (seconds)

Nu

mb

er

of

File

s

StoreEasy Deduplication Overhead (Opening Microsoft Office Files - .doc, .xls, .ppt)




Table 3. Deduplication Overhead – Opening Multiple, Small Files

Number of Files Total Time to Open All Files

Before Deduplication (Seconds)

Total Time to Open All Files After Deduplication

(Seconds) 1 1.312 1.321

22 7.171 7.702

83 27.765 28.171

125 35.638 36.799


In all cases, the time taken to open a group of files after deduplication took no longer than 3% more compared to the same files before they were deduplicated.

Users should see little to no overhead when opening smaller, deduplicated Office files. Processor utilization remained relatively low while rehydration of the office files occurred, serving as

another layer of transparency to users concerned about the impact of rehydrating to other server workloads.

The second way ESG Lab tested the potential overhead of rehydrating deduplicated files was by copying large VHDs from one volume to another volume by using the xcopy command. A scenario in which this would fit would be having a library of VHDs and copying one of them to provision a new VM on a Hyper-V host. Reasonable rehydration rates with minimal resource usage are crucial to any business looking to benefit from StoreEasy deduplication technology. ESG Lab monitored CPU utilization and elapsed time to characterize the overhead of rehydrating larger, more fragmented files. The results are shown in Figure 18 and Table 4.

Figure 18. Deduplication Overhead – Copying Large Files

Table 4. Deduplication Overhead – Opening VHDs

Number of VHDs

Consumed Capacity (GBs)

Total Time to Copy VHD Before Deduplication (Seconds)

Total Time to Copy VHD After Deduplication (Seconds)

1 10.7 19.5 25.3

1 16.5 29.8 39.7

6 81.8 164 209.9

0 25 50 75 100 125 150 175 200 225

81.8

16.5

10.7

Elapsed Time to Transfer VHD(s) (seconds)

Cap

acit

y C

on

sum

ed

on

Dis

k (G

Bs)

StoreEasy Deduplication Overhead (Copying VHDs)





ESG Lab copied VHD files contained within the volume that was deduplicated at a rate of 98%. In all cases, the time taken to copy deduplicated VHDs from one volume to another volume took around

30% longer than the VHD(s) prior to deduplication. This was true regardless of the number of VHDs. Due to the fragmentation of the highly deduplicated VHD files, CPU utilization increased modestly by 20%

when rehydrating those files.

Why This Matters

Data growth continues to challenge IT organizations. ESG research proves this point; respondents to ESG’s annual IT spending surveys for the past three years have listed managing data growth as one of the top five IT priorities. The costs of storing and managing duplicate file data can stress capital and operational budgets unnecessarily. Deduplication appliances are available, but they add another point of management to the IT administrator’s task list, another line item to the purchase order, and additional CAPEX and OPEX costs.

ESG Lab validated the deduplication capabilities of the HP StoreEasy 5000 and feels that the combination of capacity savings and rehydration rates offer a great option to any size business looking to consolidate data in a single cost-effective solution. Deduplication was easy to configure and monitor through the Server Manager interface. ESG Lab witnessed capacity savings of up to 98% with a highly deduplicateable volume containing OS images. Combined with the capacity savings offered by StoreEasy, the transparency of rehydration to users was impressive. For smaller, deduplicated files, the overhead to rehydrate when opening an Office file went unnoticed. For larger files, rehydration took about 30% longer, but with a 98% deduplication rate, ESG Lab feels that 30% overhead is more than reasonable.



ESG Lab Validation Highlights

In less than 90 minutes, ESG Lab installed and configured an HP StoreEasy 5000 Storage system and had configured an NFS-shared datastore to house virtual machines in an existing vSphere 5.1 environment.

The solution installed quickly and easily using excellent wizards. More than 25TB of usable storage was provisioned and shared out via NFS for VMware and SMB/CIFS for Windows clients in minutes, including configuring those clients to access the shares.

ESG Lab validated that the improvements to the SMB protocol will enable organizations to deploy applications like SQL Server 2012 without a need for concern for availability or performance.

Excellent performance was achieved via the HP StoreEasy 5000, an extremely cost-effective solution that leverages the new benefits of the SMB 3.0 protocol, consolidation through VMware vSphere, and easy provisioning.

In a VMware vSphere environment, the HP StoreEasy 5000 showed excellent NFS performance as well, scaling nearly linearly with mixed business-critical workloads including more than 12,000 Exchange mailboxes on virtual machines using storage provisioned from an NFS-mounted datastore.

ESG Lab validated the deduplication capabilities of the HP StoreEasy 5000 and feels that the combination of capacity savings and rehydration rates offer a great option to any size business looking to consolidate data in a single cost-effective solution.

Deduplication was easy to configure and monitor through StoreEasy’s Server Manager interface. ESG Lab witnessed capacity savings of up to 98% with a highly deduplicateable volume containing OS

images. Combined with the capacity savings offered by StoreEasy, the transparency of rehydration to users was impressive. For smaller, deduplicated files, the overhead to rehydrate when opening an Office file went unnoticed.

Issues to Consider

Default server BIOS, operating system, and application settings were used during ESG Lab testing. As expected, after any testing of this magnitude, analysis of the results indicates that tuning would most likely yield slightly higher absolute results. Given that the goal of this report was not to generate a big number, ESG Lab is confident that the results presented in this report meet the objective of demonstrating the new functionality and capabilities of the HP StoreEasy 5000, as well as performance and scalability benefits of business critical applications and workloads.



The Bigger Truth

Data volumes are growing at unprecedented rates and that fact keeps IT administrators awake at night. Combined with significant downward pressure on IT budgets, this unmitigated data growth stretches IT budgets and resources to the breaking point. IT is being forced to rethink its approach to grappling with this explosion of data. Rising end user performance and stability expectations have exacerbated these challenges significantly.

Server virtualization has been proven to deliver significant business value, reducing both OPEX and CAPEX and keeping virtualization at the top of the IT priority list. But while customers continue to enjoy efficiency improvements, they want to move past the initial consolidation gains and virtualize more mission-critical applications. However, the workload aggregation of virtual server environments can create an “I/O blender” effect, with multiple types and sizes of workloads overloading disk spindles, stressing networks, and consuming server processing resources. These challenges can affect application performance, often preventing expansion of virtual environments and limiting its benefits.

Organizations are hungry for systems and tools to become more efficient and smart about storage management. Hewlett-Packard has synergistically combined the HP StoreEasy 5000 with Microsoft’s Windows Storage Server 2012 to produce a solution that is truly greater than the sum of its parts. The result is a highly available, highly scalable, small footprint, easy-to-order/deploy/manage offering with robust performance.

ESG Lab testing has demonstrated just how groundbreaking the release of the StoreEasy 5000. The new Server Manager interface offered a refreshing way to configure and manage storage with easy-to-use wizards, and intuitive navigation. The improved SMB protocol provided flexibility to organizations while still meeting strict performance requirements and NFS 4.1 provided robust, easy to configure, highly available storage for a VMware vSphere environment. ESG Lab proved that business critical applications, such as SQL Server 2012 and Exchange Server 2010, can not only meet performance expectations, but exceed them, in a consolidated virtual server environment powered by the HP StoreEasy 5000.

ESG Lab testing showed that StoreEasy 5000 contained a mature set of software tools combined with sophisticated hardware. In less than 90 minutes, ESG Lab was able to completely implement the StoreEasy 5000, including clustered controller setup, provisioning, configuring and sharing out 25TB of storage over multiple protocols.

Given the prevailing notion that IT should be able to do more with less, HP’s StoreEasy 5000 lets IT save money by avoiding the expense and bother of supporting a complex, dedicated storage area network. Instead, the StoreEasy 5000 leverages existing network infrastructure and common HP server and storage components to deliver unified storage—an uncomplicated and less expensive option.

The HP StoreEasy 5000 delivered exactly what HP intended: a viable turnkey solution for the midrange file storage space, with impressive performance and a raft of features that appeal to harried IT administrators. The resultant service delivery to end customers is competitive but cost effective, while the storage efficiency, agility, and transparency of new and improved features make the StoreEasy 5000 high-performing and cost-effective. ESG Lab believes that IT managers in mid-sized organizations looking for a robust, high performance, highly available file services solution would be wise to consider the HP StoreEasy 5000.



Appendix

Table 5. ESG Lab Test Bed

Useability, Functionality, and SMB Testing

HP StoreEasy 5530 (B7E07A)

Operating System: Windows Storage Server 2012 Drives: 36x 900GB 10K SFF SAS Volume Count: 2 (one for each node) Capacity: 32.4 TB Raw, 25.39 TB Usable RAID 6 protected

Physical Servers

Type:2x HP ProLiant ML150 G6 Processor: 2x Quad core 2.13GHz Intel Xeon X5506 Memory: 96GB Operating System: Windows Server 2012 OS Build: 9200

Virtual Machines

Virtual Machine Count: 4 vCPUs: 2 RAM: 4GB Operating System: Windows Server 2012 OS Build: 9200

Table 6. ESG Lab Deduplication Powershell Commands

Powershell Command Description

Enable-DedupVolume <volume>: This command enables deduplication on a specified

volume

Set-DedupVolume -<volume>: -MinimumFileAgeDays 0

By default, deduplication optimizes files that are modified at least 5 days in the past. This command allows for the

optimization of all files regardless of creation and modification timestamp.

Set-DedupSchedule * -Enabled:$false This command ensures that the dataset in the volume

does not get deduped prior to measurement.

Start-DedupJob -<volume>: -Type Optimization This command runs deduplication in throughput mode.

Start-DedupJob -<volume>: -Type Optimization –Memory 25 -StopWhenSystemBusy

This command runs deduplication in background mode.

20 Asylum Street | Milford, MA 01757 | Tel: 508.482.0188 Fax: 508.482.0218 | www.enterprisestrategygroup.com

lab validation report -...

Documents