Lab Validation

Report
Fujitsu ETERNUS DX8000

Large Scale Storage Consolidation for the Enterprise

By Tony Palmer with Mark Peters

September 2009

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 2

Contents
Introduction.............................................................................................................................................3
Background......................................................................................................................................................... 3
Fujitsu ETERNUS DX ............................................................................................................................................ 4
ESG Lab Validation ...................................................................................................................................5
Ease of Deployment and Flexible Management ................................................................................................... 5
Eco-mode ......................................................................................................................................................... 10
Fujitsu ETERNUS DX Data Encryption................................................................................................................. 13
Performance and Scalability .............................................................................................................................. 16
SPC-1 Results .................................................................................................................................................... 18
Availability and Business Continuity .................................................................................................................. 20
VMware Site Recovery Manager ....................................................................................................................... 21
ESG Lab Validation Highlights................................................................................................................. 25
Issues to Consider ..................................................................................................................................25
The Bigger Truth ....................................................................................................................................26
Appendix ...............................................................................................................................................27

ESG Lab Reports

The goal of ESG Lab reports is to educate IT professionals about emerging technologies and products in the
storage, data management and information security industries. 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 Fujitsu.

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, whic h 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 applic able, criminal prosecution. Should
you have any questions, please contact ESG Client Relations at (508) 482.0188.

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 3

Introduction
The Fujitsu ETERNUS DX8000 storage system is an extremely scalable enterprise-class storage platform that
combines massive scalability (up to 2.7 petabytes) with data availability, security, and operational efficiency. This
ESG Lab Report documents the results of hands-on testing of the DX8000 with a focus on performance and capacity
scalability, availability, data mobility, power efficiency, and security.

Background
Organizations of all sizes are struggling to meet the conflicting challenges associated with macroeconomic global
financial uncertainty and micro-level information storage growth and complexity. As users continue to struggle to
manage increasing volumes of data, they are also tasked to do so with constrained—or even reduced—resources.
Consequently, users are more willing than ever—even compelled, in certain cases—to look for new ways to address
their needs.
Figure 1. Top Storage Challenges

In general, what are your organization’s greatest challenges with respect to its
storage environment? (Percent of respondents, N=504, multiple responses accepted)

Keeping pace with overall data growth 36%

Storage system costs 35%

Securing confidential data 31%

Need to improve backup and recovery processes 31%

Power and cooling costs 28%

General increase in complexity of IT environment, number of

servers, etc. 28%

Running out of physical space 26%

Storage requirements/complexity related to specific

applications (e.g., e-mail, database, home-grown 25%
applications, etc.)

Insufficient capital budget 24%

0% 10% 20% 30% 40%

Source: Enterprise Strategy Group, 2009.

As recent ESG research of the enterprise IT market shows (see Figure 1), users are heavily focused on the joint
challenges of physical resource constraints (as evidenced by data growth, capital cost, power and cooling expenses,
and floor space issues) and management resource limitations (as evidenced by the need to improve backup and
recovery processes as well as addressing complexity in numerous areas).1

1
Source: ESG Research Report, 2008 Enterprise Storage Systems Survey, November 2008.

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 4

Fujitsu ETERNUS DX
Fujitsu has focused on these customer trends and priorities with its ETERNUS DX family of disk storage systems.
ETERNUS DX is designed to address the most crucial challenges faced by storage and IT managers.
Figure 2. The Fujitsu ETERNUS DX Family

The ETERNUS DX8000 is a massively scalable enterprise storage consolidation platform supporting both open
systems and mainframe connectivity. In addition to massive scalability, the DX8000 offers advanced storage
attributes and functionality that, though derived from Fujitsu’s mainframe roots, is attuned to the needs of today’s
enterprise:
Capacity and Performance Scalability – ETERNUS boasts the industry’s largest storage capacity (over 2.7
petabytes) and 2.83GHz quad core processors to significantly increase performance over previous
generations.
Data Integrity – ETERNUS DX offers data block checking and RAID6 (double parity) support.
Enhanced Security - Integrated hardware-based data encryption, both ‘on the fly’ for remote copy
functions and ‘at rest’ to secure data on drives removed from service, secures users’ data.
Data Protection and Availability - A variety of high-speed internal and external volume copy functions
provide high availability and business continuance.
Flexible Management and TCO Reduction - Capacity on Demand (CoD), online data migration, and thin
provisioning capacity virtualization enable better utilization of resources with no disruption to operations.
Green IT – Power and cooling requirements are reduced with Eco-mode using MAID technology and low-
power, high performance Enterprise SSD support.
The balance of this report presents the results of hands on testing of the Fujitsu ETERNUS DX disk storage systems
with an eye toward validating the performance, scalability, and reliability of the platform while examining its
advanced availability, environmental, and security features.

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 5

ESG Lab Validation

ESG Lab performed hands-on evaluation and testing of the Fujitsu DX8000 at Fujitsu’s lab in Sunnyvale, CA. Shown
in Figure 3, testing utilized Fujitsu storage systems connected via Fibre Channel and iSCSI to physical and virtual
servers. Testing was designed and executed using industry standard tools and methodologies to validate
performance, reliability, operational efficiency, security, and availability as well as integration with VMware Site
Recovery Manager software.2
Figure 3. The Fujitsu ETERNUS DX Test Bed

Ease of Deployment and Flexible Management

ESG Lab first examined the management interface of the ETERNUS DX, evaluating the relative ease with which
volumes could be created, managed, and manipulated by a storage administrator. Fujitsu offers Storage Cruiser, a
full featured client-server management application, as well as ETERNUSmgr aid, providing basic storage
provisioning and performance management functionality.

ESG Lab Testing

The first step was to launch the Storage Cruiser Configuration Navigator, seen in Figure 4. The column on the left
shows icons for all arrays under management at the top with disk level detail beneath. The main panel is a visual
representation of the selected system. ESG Lab right-clicked in the main panel and selected ‘RAIDGroup Operation.’

2
Configuration details can be found in the Appendix.

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 6

Figure 4. Storage Cruiser Configuration Navigator

The RAIDGroup Operation screen, seen in Figure 5, shows all disks in the system and is color-coded for clarity.
Drives in RAID groups are identified by color, white designates hot spares, and dark gray indicates drives available
for assignment. ESG Lab selected eight available drives and created a RAID6 group.
Figure 5. Creating a RAID Group

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 7

Next, a 10 GB logical volume was created from the available capacity in the newly created RAID group. ESG Lab then
created an ‘affinity group,’ which is a key part of Fujitsu’s volume masking technology and assigned a pair of
ETERNUS DX fibre channel ports to it. 3 The 10 GB volume was added to the new affinity group and presented to an
attached server by assigning the host WWNs to the affinity group. The last step was to perform a rescan in
Windows disk administrator and format the new 10 GB volume. Start to finish, the entire volume creation and
assignment process took about 10 minutes. The Resource Coordinator Tool (not tested by ESG) enables access path
configuration, which gives the ability to configure end-to-end connectivity from ETERNUS DX fibre channel port to a
host HBA in one operation.
ESG Lab tested the ETERNUS DX platform’s ability to migrate volumes online, with a system actively reading and
writing to the volume. ESG Lab simulated an active application using the industry standard Iometer benchmarking
utility.4 First, Iometer was started on a server attached to the ETERNUS DX under test. A continuous OLTP database
workload was then run against a 10 GB volume mounted from the DX. While this simulated application workload
was reading and writing to the 10 GB volume, ESG lab used ETERNUS Manager to select the volume for migration,
seen in Figure 6.
Figure 6. Online RAID group Migration Setup: Select Source

3
Affinity groups contain unique host identifiers (Fibre Channel WWNs or iSCSI initiator IQN), WWNs of the ETERNUS fibre channel ports and
the volumes to be assigned to a host or group of hosts.
4
Iometer is an open-source I/O measurement and characterization utility, available for download at: http://www.iometer.org/

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 8

Next, the destination RAID group was selected and available space was allocated, as shown in Figure 7. Finally, ESG
Lab clicked ‘Set’ to execute the migration. ETERNUS Manager requested confirmation, displaying the source and
destination RAID groups. Upon clicking ‘OK,’ the volume migration began. A quick look at the server confirmed that
Iometer was still running with no disruption as the migration took place.
Figure 7. Online RAID group Migration Setup: Configure Target

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 9

Figure 8. Migration Complete: New Space Available

Within a few minutes, the volume migration completed and ESG Lab then expanded the volume using Microsoft
Disk Administrator (see Figure 8). A final examination of Iometer showed that the application was still reading and
writing to the drive with no errors.

Why This Matters

Storage deployments are growing in capacity and complexity within organizations of all sizes while IT managers are
increasingly being asked to manage this growing storage capacity with stagnant, or shrinking, budgets and staffing.
Efficient deployment and management of storage is essential.
Fujitsu ETERNUS DX provides a straightforward management interface that offers advanced storage functionality
with minimal complexity. ESG Lab was able to provision new storage and present it to a server in less than 10
minutes. Data migrations were just as easy and completely non-disruptive, which enables administrators to provide
better service levels to applications and users while making it easier to manage the system over its lifetime, helping
further reduce costs and do more with less.

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 10

Eco-mode
Eco-mode is Fujitsu’s term for power consumption reduction using MAID technology. MAID, an acronym for
"Massive Array of Idle Disks," reduces power consumption by spinning down less frequently accessed disk drives
when not in use. Eco-mode can be set for any type of disk or RAID group in the array. After 30 minutes idle time,
drives are spun down, reducing power consumption. Read or write access to any volume on any drive in a spun
down RAID group will re-activate the drives. Re-activation can take anywhere from seconds to up to three minutes,
depending on the drive type and configuration.
Figure 9 shows the timeline of a use case for Eco-mode. In this instance, the customer uses a pool of large capacity
SATA disks as backup to disk targets. The drives are scheduled to spin up just before midnight, when nightly backup
jobs kick off. Thirty minutes after the last backup job completes, the drives spin down. The drives remain idle until
the next scheduled spin up or, in this case, a user requests a restore. At 9:01AM, the backup administrator initiates
the restore from the backup application. The drives spin up automatically and the restore runs. Thirty minutes after
the restore job completes, the drives spin down again.
Figure 9. Fujitsu ETERNUS DX Eco-mode

In this scenario, the backup to disk pool remains spun down for about 80% of the time, representing significant
power savings.

ESG Lab Testing

First, ESG Lab enabled Eco-mode for RAID group 0x005, which contained a volume presented to a Fibre Channel-
attached Windows server. Eco-mode was enabled by simply setting the Eco-mode checkbox and clicking ‘Set,’ as
seen in Figure 10.

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 11

Figure 10. Configuring Fujitsu ETERNUS DX Eco-mode

There is a drop down menu to select a schedule for disk active times if one is required. As seen in Figure 10, RAID
group 0x005 was active at the time Eco-mode was enabled. ESG Lab waited 30 minutes after setting Eco-mode and
checked back in using ETERNUS Manager. Figure 11 shows that the RAID group was reported as idle at this point.

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 12

Figure 11. Verifying an Idle RAID Group

Next, ESG Lab issued the ‘dd’ command on the Windows server. As shown in Figure 12, the response time after the
first command was issued was 11,078 ms, or just over 11 seconds. The second command to the volume returned a
response in 16ms, exactly as expected as the volume was spun back up. To help put this into context, the Fujitsu
design minimizes the power footprint by staging the spin-up of drives. This reduces the maximum power demand
that would result from powering up all drives simultaneously. While simultaneous spin-up would result in faster
initial response times, it would also require higher peak power demands on the customer’s environment.
Figure 12. Eco-mode in Action

Fujitsu is conservative in its estimates and states that drives can take between one and three minutes to re-activate
after spinning down. This depends heavily on the drive and RAID configuration; it will certainly take longer for a
RAID6 group made up of a large number of SATA drives to re-activate than a RAID-1 Fibre Channel drive pair.

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 13

Why This Matters

Energy costs are rising steadily and some metropolitan data centers in the United States are already reaching the
limits of local utility providers. ESG research indicates that 71% of enterprises surveyed have formal initiatives and
programs in place to reduce overall data center power and cooling requirements. 5 At odds with this is the
incredible rate of data growth occurring globally. As more and more data is created and retained, spinning disk
drives represent a significant portion of energy consumed and heat generated in the data center.
ESG Lab has confirmed that Fujitsu ETERNUS DX Eco-mode can be use to easily and effectively reduce power and
cooling requirements by spinning down idle drives when not in use. In environments making use of large numbers
of SATA drives for backup to disk or archiving, this represents a significant savings opportunity.

Fujitsu ETERNUS DX Data Encryption

Thanks to information protection legislation—including Sarbanes Oxley in the United States and Japan's Financial
Instruments and Exchange Law—information security has gained greater significance in the enterprise. These laws
require that access to stored data is restricted to authorized users and unauthorized or accidental access to
sensitive information is blocked. ETERNUS DX disk storage systems provide on-disk data encryption to address this
requirement. Data can be automatically encrypted inside ETERNUS DX storage systems using either industry
standard 128bit AES encryption or a proprietary Fujitsu algorithm. This ensures that data is protected even if the
physical drives are removed from the ETERNUS DX system. The data at rest encryption is performed at the logical
volume level and is completely transparent to users with no access control or key management required—the
encryption keys are tied to the RAID controllers in each individual ETERNUS DX system.
Figure 13. Fujitsu ETERNUS DX Data Encryption

5
Source: ESG Research Report, Global Green IT Priorities: Beyond Data Center Power and Cooling, November 2008.

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 14

ESG Lab Testing

ESG Lab tested data encryption in a Fujitsu ETERNUS DX440. Using the ETERNUS Manager console, ‘Convert
Encryption Volume’ was selected from the RAID management menu. This brought up a list of logical volumes with a
checkbox next to all unencrypted volumes. ETERNUS Manager allows selection of individual volumes or a range, as
seen in Figure 14.
Figure 14. Convert Encryption Volume

After selecting volume 0x0030, a volume visible to a test server containing a group of copied files, ESG Lab began
the encryption process. Encryption status was visible directly from ETERNUS Manager. Once encryption was
complete, ESG Lab confirmed that the files on the drive were still visible from the test server. Next, Fujitsu removed
all the drives in the RAID group containing volume 0x0030 from the DX440 system and physically installed them in
another DX440.

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 15

Figure 15. Moving Drives to a Different Array

As seen in Figure 15, the second DX440 was unable to read any of the encrypted volumes, returning errors for each
encrypted volume on the drives. The implementation of encryption in the ETERNUS DX400 and DX8000 is identical,
so testing with the DX440 provides a user experience and result exactly the same as would be achieved with an
DX8000.

Why This Matters

Widely publicized data breaches, privacy laws, and boardroom jitters are driving a behavioral shift towards data
encryption. ESG research indicates that lost or stolen IT assets are a major source of anxiety for enterprises
concerned with confidential data vulnerability. A significant percentage of these organizations believe that
encryption technology should be integrated into disk and controller hardware. 6 Should personal or financial data
become compromised, the corporation would potentially be liable for damages and heavy fines. The costs
associated with the loss or theft of personal credit card information alone has been staggering: Consider, for
example, one of the many public disclosures associated with the loss of tapes containing the personal data of 3.9
million customers. Estimating a cost of $30 to $150 per customer for notification and credit services, this single
event translates into a total cost of between one and six billion dollars.
These incidents demonstrate that the risk is real and the costs are high. Encrypting data on disk reduces risk, avoids
potentially crippling costs, and keeps the CEO out of the headlines. ESG Lab verified that the Fujitsu ETERNUS DX
platform secures data at rest with integrated AES encryption and prevents unauthorized data access after disk
drives are removed from the system.

6
Source: ESG Research Report, Protecting Confidential Data, March 2006.

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 16

Performance and Scalability

The Fujitsu DX8000 is an enterprise-class disk storage system supporting both Fujitsu GlobalServer mainframes and
UNIX/industry standard servers in large-scale commercial and scientific environments. The DX8000 was architected
to meet the high-capacity and performance requirements of large organizations while providing a foundation for
matching application requirements with appropriate classes of storage. Its platform offers non-stop operation
thanks to multiple active controllers, Cyclic CM, a fully redundant back end, and multiple local and remote data
replication options. Cyclic CM enables DX8000 disk storage systems to operate with up to eight controllers,
enabling performance and high availability to be maintained during a controller failure or scheduled maintenance.
With the Cyclic CM function, write data is duplicated to cache memory on different controllers. If one controller
fails or is taken out of service, the remaining controllers automatically reconfigure to close the gap and ensure
100% cache and controller redundancy.
The basic ETERNUS DX8000 architecture is shown in Figure 16. The architecture of the ETERNUS DX is designed to
eliminate single points of failure and to increase availability as more resources are added to the system. Servers
access the ETERNUS DX through front end channel adapters (1) which support active-active load balanced
redundant connectivity. Multiple high speed PCI-Express buses (2 and 3) are used to move control information and
data within and between Control Modules. Multiple redundant Fibre Channel loops (4) connect each disk enclosure
to multiple controllers.
Figure 16. The Fujitsu ETERNUS DX8000 Architecture

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 17

ESG Lab Testing

ESG Lab next analyzed ETERNUS DX system performance in an Exchange environment via an audit of Microsoft
ESRP 2.1 results and hands-on testing using Iometer workloads designed to simulate a Microsoft Exchange 2007
environment. Fujitsu has published a Microsoft Exchange Solution Review Program (ESRP) result for a two
controller ETERNUS2000 Model 100. 7 ESG Lab performed a series of tests against a four controller DX8400 to
simulate Microsoft Exchange 2007 workload scaling using Iometer.
ESRP is a Microsoft program designed to facilitate third party storage testing and solution publishing for Exchange
Server. The program combines a storage testing harness (JetStress) with publishing guidelines for Microsoft Gold
Certified and Storage OEM Partners. ESRP employs the Jetstress utility to create real exchange IO traffic that runs
against real exchange databases—with logging and file attachments—exactly as in the real world. The testing is
designed to measure both the performance and reliability of a given solution. The performance test runs for two
hours while the reliability test runs for 24 hours. Both tests must run without exceeding a prescribed disk latency
threshold (20 milliseconds) and a reliability test is performed to check for database and log corruption at the end of
the run. Manufacturers use the ESRP framework to test storage solutions and then submit results to Microsoft for
review. Links to approved solution results are posted on the Microsoft Exchange ESRP website.8
The published ESRP results are for a relatively small configuration supporting 2,500 users on just 14 disk drives
using Microsoft’s Heavy user profile (.384 IOPS per mailbox). ESG Lab testing was performed against a DX8400 using
14, 28, and 42 drives in a RAID 1+0 configuration. Iometer was used to generate a workload similar to the IO
generated by JetStress to provide a quick estimate of the number of Exchange 2007 users that the system could
support.
Figure 17. ESG Lab Exchange Simulation Testing

40,000

35,000

30,000
Exchange Mailboxes

25,000

20,000

15,000

10,000

5,000

0
14 28 42 56 70 84 98 112

Number of Drives
ESG Lab Projected ESG Lab Tested
Source: Enterprise Strategy Group, 2009.
ESG lab observed a scaling factor of approximately 92% running the Exchange workloads on successively larger
groups of drives. Figure 17 shows the projected scalability of the ETERNUS DX platform’s support for Exchange
mailboxes using Microsoft’s Very Heavy user profile of .5 IOPs per mailbox. Scaling was nearly linear and response

7
http://www.fujitsu.com/downloads/STRSYS/system/eternus2000-esrpv.pdf
8
http://technet.microsoft.com/en-us/exchange/bb412164.aspx

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 18

times were about 15 ms for all tests. Table 1 shows the results of Iometer testing and estimated supportable
Exchange mailboxes using Microsoft’s Very Heavy user profile.
Table 1: Iometer Test Results and Projection

Number of Drives Exchange DB IOPS Number of Exchange Mailboxes

14 3,094 6,187
28 5,955 11,909
42 8,587 17,174
112 (projected) 18,907 (projected) 37,813 (projected)

What the Numbers Mean

The DX8400 provided sufficient IO to support 17,174 Exchange users with just 42 disk drives.
IO scaled nearly linearly as drives were added, with minimal differences in response time. This is highly
desirable as disk response time has a direct impact on the user experience in an Exchange environment.
This simulation only projects performance out to 112 drives. The DX8400 scales to 1072 drives while the
DX8700 scales to 2,760 drives. ESG Lab feels strongly, based on hands-on testing and audit of published test
results, the DX8000 platform can service much higher numbers of Exchange users.

SPC-1 Results
ESG Lab audited Fujitsu’s published results of the SPC-1 application-level industry standard benchmark suite
maintained by the Storage Performance Council. SPC-1 testing generates a single workload designed to emulate
the typical functions of transaction-oriented, real-world database applications. Transaction-oriented applications
are generally characterized by random IO and generate both queries (reads) and updates (writes). Examples of
those types of applications include OLTP, database operations, and mail server implementations.
Figure 18. Fujitsu DX8400 SPC-1 Results
30
100%

25
Response Time (ms)

20

15

10 109,246 SPC-1 IOPS

3.84 ms Response Time
5
95%
50% 80% 90%
0 10%
0 20,000 40,000 60,000 80,000 100,000 120,000 140,000

IO Requests Per Second

Source: Enterprise Strategy Group, 2009.

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 19

As seen in Figure 18, Fujitsu has published an excellent result of 109,246 SPC-1 IO requests per second at 95% load
with an average response time of less than 3.84 milliseconds. Response time is an extremely important component
of SPC results, as this is the delay that an application will experience (and pass on to users) when a storage system
is stressed to its limits. A system servicing 109,246 SPC-1 IO requests per second with a response time of 3.84 ms is
exceptional.
The SPC-1 benchmark synthesizes a community of users running against storage that is organized logically as would
be encountered in a real-world application. SPC is one of the few benchmarks in the industry that helps to deliver
value in the area of real world performance characterization. SPC results can be roughly mapped by users into
easily understood metrics. For a credit card database system, for instance, it might be the number of credit card
authorizations that can be executed per second.
It should be noted that the numbers published in these reports are for the previous generation Fujitsu ETERNUS DX
subsystems. As of this writing, Fujitsu is working toward the release of updated benchmarks which should present
higher performance numbers than currently published.
SPC-1 results are audited by the Storage Performance Council and peer reviewed to ensure consistency. Executive
Summary and Full Disclosure Reports (FDRs) for each SPC benchmark result are publicly available for download and
review.9 While this can be useful for comparison between vendors, it is important to note that not all vendors
participate and publish results. ESG Lab applauds Fujitsu’s participation in this peer-reviewed program and hopes
that it will encourage other vendors of enterprise class storage systems to participate.

Why This Matters

ESG research indicates that performance is a key concern when deploying mission critical applications in a highly
consolidated environment. With multiple application servers relying on a shared storage infrastructure, there is a
worry that performance requirements can‘t be met affordably. As a matter of fact, 51% of ESG survey respondents
reported that performance was their top concern, followed closely by capital and operational costs. 10
Through careful examination of ESRP and SPC-1 results, combined with hands-on testing, ESG Lab has verified that
Fujitsu ETERNUS DX8000 disk storage systems can be deployed to cost-effectively provide predictably scalable
enterprise-class storage for Exchange and OLTP environments of all sizes with excellent price-performance, as
evidenced by its excellent $16.12/SPC-1 IOP, which is the best published result for a mainframe-ready enterprise
disk based storage system. As enterprise environments continue to grow in size and complexity, the DX8000 can
clearly provide a platform for affordable performance and capacity scalability.

9
http://www.storageperformance.org/results/benchmark_results_spc1
10
Source: ESG Research Report, The Impact of Server Virtualization on Storage, December 2007.

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 20

Availability and Business Continuity

Fujitsu ETERNUS DX disk storage systems offer a broad range of data and volume protection options that enable
organizations to provide highly available storage to their users within a single system and across multiple data
centers. ETERNUS DX provides extensive, enterprise class, in-system data integrity and availability assurance. For
example, the Data Block Guard feature adds check codes to data and verifies the data at the Channel Adapter and
Device Adapter to ensure operational accuracy and end-to-end storage integrity. Data availability is ensured via
component redundancy, hot swap, failure isolation, multi-path access as well as intra- and inter-system volume
copy services.

ESG Lab Testing

Due to the similarity in design and implementation of availability features across the ETERNUS DX line, ESG Lab was
able to test system hardware availability features against an ETERNUS DX440 and achieve the same results as would
be obtained testing against an ETERNUS DX8000. While running a continuous Iometer workload against a volume
presented to a virtual Windows 2003 server in an attached VMware ESX host, multiple errors were injected. A disk
drive, power connection, Fibre Channel cable, and CM (Controller Module) were all removed from the system, one
at a time, to simulate multiple failures on a running system. First, a drive was removed from the RAID group
housing the volume under test. Figure 19 shows the ETERNUS Manager hot maintenance screen displaying the RAID
group after the drive was removed from service. The hot spare drive activated automatically and the RAID group
began to rebuild.
Figure 19. The ETERNUS Manager Hot Maintenance Screen

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 21

It’s important to note that the procedure for replacing the drive is clearly presented by ETERNUS Manager, giving
customer service personnel unambiguous indication of the drive to be replaced and the methodology for non-
disruptively doing so.
Next, ESG Lab pulled a power cable, removed two of the four Fibre Channel connections to the ESX server, and
disabled one of the two main controller modules in the system, pausing to examine ETERNUS Manager between
each event. After each event, the administrator received an e-mail detailing the fault and a call was placed to
Fujitsu support. Iometer continued to run through every event, with no errors and only a momentary pause in IO
when the controller module was forced down.

VMware Site Recovery Manager

The test bed used for VMware SRM is summarized in Figure 20. Disaster recovery plans were created to manage
the failover, and failback, of web and wiki servers running in the primary data center. Documents were uploaded to
the web server and edited while the site failover was performed.
Multiple servers were deployed in the simulated primary and secondary data centers running our test applications
as well VMware Infrastructure Manager and VMware vCenter Site Recovery Manager.
An ETERNUS DX8400 system was deployed as the primary system and an ETERNUS DX440 was deployed as the
secondary. Gigabit Ethernet iSCSI ports on the two storage systems were used to connect the two data centers to a
simulated wide area network directly with hardware-based encryption and WAN optimization.11
Figure 20. The VMware SRM Test Bed

VMware ESXi version 3.5, update 4, was used to deploy virtual machines in each of the data centers. All of the
virtual machines ran the Windows 2003 Enterprise x64, SP1 operating system. The DX8400 system ran firmware
version V11L71 and the DX440 ran firmware version V20L15. ETERNUS SF Advanced Copy Manager was used to
replicate data volumes between systems.

11
See the Appendix for more configuration details.

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 22

ESG Lab Testing

ESG Lab testing began with a review of a pre-configured VMware vCenter SRM disaster recovery plan. The steps
required to configure a VMware vCenter SRM disaster recovery plan are:
1. Install VMware vCenter SRM software on a server in each data center
2. Install the ETERNUS SF Site Replication Adaptor on the VMware vCenter SRM server in each data center
3. Launch the VMware vCenter SRM GUI to create a disaster recovery plan
4. Provide the IP address of the key components in the primary data center, including the ETERNUS system
5. Provide a list of virtual machines to fail over
VMware vCenter SRM does the rest; including discovering the virtual machine resources (memory, CPU, network,
and drive resources) needed to automatically recreate a working application environment after a disaster.
VMware recovery plans were run during ESG Lab testing to fail over a web server application from a primary data
center to a secondary data center. A failover with both data centers up and operational was performed to confirm
that the virtual machines protected by a VMware recovery plan (and the applications running in those machines)
could be successfully restarted at the secondary data center with no data loss. ESG Lab ran the site failover
immediately after an administrator added a document to a page on the web server and viewed the page to confirm
the new document was up on the site. Figure 21 shows the progress of the recovery moments after the Run button
was clicked. ESG Lab noted that progress was very easy to follow. Successfully completed recoveries are depicted
in green, the currently executing steps are blue, and any failed steps are shown in red.
Figure 21. Running a Disaster Recovery Plan

The fully automated recovery completed in less than five minutes. Three minutes after the failover had completed,
the virtual machine was booted and running, as seen in Figure 22.

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 23

Figure 22. VM Ready to Run at Secondary Site

Figure 23 shows the website up and running on a virtual machine in the secondary system.
Figure 23. Applications Recovered and Running at Secondary Site

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 24

Once ESG Lab confirmed that the web server was up and that all documents and pages from the primary system
were online and complete, the systems and applications were failed back to the primary data center. The VMware
infrastructure manager console at the primary site was used to define a new recovery plan using the wizard-driven
process described earlier in this report (see page 22). Except for the swapping of IP addresses for the VMware and
storage resources at the secondary data center, the process was exactly the same.
Less than thirty minutes after getting started with the definition of a failback plan, the website was back up and
running at the primary data center.

Why This Matters

Recovering from a disaster using traditional backup software methods can take days. Disk-based remote mirroring
can be used to cut the recovery time to minutes, but traditional solutions which rely on scripts and manual
operations can be complex and error prone. When sprinklers are running and cell phones are ringing, complex
manual operations are the last thing an IT staffer needs to deal with.
ESG Lab has verified that Fujitsu ETERNUS DX disk storage systems running ETERNUS SF Advanced Copy Manager,
in conjunction with VMware SRM, can be used to automate the recovery of critical applications after a disaster.
Less than ten minutes after a simulated failure in a primary data center, applications were up and running at a
remote recovery site with zero data loss.
Restoring operations in a rebuilt or new data center is typically the very last step in a disaster recovery process. ESG
Lab confirmed that the same wizard-driven process used for failover can also be used for failback.

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 25

ESG Lab Validation Highlights

 ESG Lab was able to provision new storage and present it to an application server in less than 10 minutes.
 Fujitsu ETERNUS DX Eco-mode was used to easily and effectively reduce power and cooling requirements
by spinning down idle drives when not in use.
 ESG Lab verified that the Fujitsu ETERNUS DX platform secures data at rest with integrated AES encryption
and prevents unauthorized data access after disk drives are removed from the system.
 Fujitsu ETERNUS DX disk storage systems can be deployed to cost-effectively provide predictably scalable
enterprise-class storage for Exchange and OLTP environments with excellent price-performance.
 Less than ten minutes after a simulated failure in a primary data center, applications were up and running
at a remote recovery site with zero data loss using ETERNUS SF Advanced Copy Manager in conjunction
with VMware SRM.

Issues to Consider
 While ESG Lab found management of the ETERNUS DX platform to be robust and stable, there are multiple
management interfaces: a client server application and a web interface to manage each system. There is
some overlap, but in some areas, these apps provide a unique subset of functionality. Ideally, ESG would
like to see these management apps consolidated or at least merged, such that all functionality is available
on whichever app an administrator chooses to utilize.

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 26

The Bigger Truth

Enterprise IT departments are being challenged to address accelerating information storage growth and complexity
without compromising performance or availability. As users continue to struggle to manage increasing volumes of
data, they are also tasked to do so with constrained—or even reduced—resources. Fujitsu is delivering a level of
capability and scalability in the ETERNUS DX that is compelling and extremely valuable to end-users.
The ETERNUS DX8000 is a highly available, flexible, and massively scalable disk storage system designed to enable
enterprises to quickly and efficiently deliver storage services to their users and customers. The ETERNUS DX
platform provides many capabilities that individually provide great value; but when these capabilities are
combined, the result is true enterprise class storage.
ESG Lab found the ETERNUS DX8000 to be easy to manage while providing impressive performance for critical
applications like Microsoft Exchange and OLTP databases. The ETERNUS DX8000 proved capable of scaling to
support more than 37,000 Exchange mailboxes with consistently crisp response times. SPC-1 results were similarly
impressive, with the ETERNUS DX8000 providing 109,246 SPC-1 IOPS with a 3.84 ms average response time.
Enterprise class attributes, as implemented in the ETERNUS DX platform, provide data integrity, hardware
reliability, and high availability options required by enterprises as well as functions not found in all enterprise class
arrays, such as Eco-mode power reduction technology as well as hardware-based encryption for remote replication
and disk volumes.
ETERNUS DX also demonstrated robust integration with VMware SRM for easy to use, automated site failover and
recovery. ESG Lab used ETERNUS SF Advanced Copy Manager remote replication technology to support a VMware
SRM deployment and failed over from one site to another in minutes.
Fujitsu ETERNUS DX uses the same architecture and software across its entire storage range to provide
demonstrable performance, scalability, and reliability benefits. The ETERNUS DX8000 provides large-scale storage
consolidation, and a range of capabilities that are everything users have come to expect from an enterprise class
array in a robust, mature package. Fujitsu has both breadth and depth in the storage business. Decades of R&D and
market success have translated into an offering capable of supporting the most critical systems for the largest
enterprises.

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

 Lab Validation: Fujitsu ETERNUS DX8000 27

Appendix
Table 2. Test Configuration

Fujitsu ETERNUS DX440 –

60x 146 GB FC drives
Firmware V20L15
8x 4 GB FC connections
4x GbE iSCSI host connections
Fujitsu ETERNUS DX8400 –
52x 146 GB FC drives
Firmware V11L71
16x 4 GB FC connections
4x GbE iSCSI host connections
2x Brocade 4100 32 port 4 Gb/sec FC switches Firmware Fabric OS v6.1.1
VMware ESXi v. 3.5 U4 Server –
1x PRIMERGY RX600 S4 Guest Operating System:
4x Xeon 2.4GHz Quad core CPU MS Windows Server 2003 SP1 x64
64 GB RAM
4x QLogic QLA2342 4 Gb/sec FC HBA
Physical Windows Server
1x PRIMERGY RX300-S3 Operating System:
1x 2.33GHz Quad core CPU Xeon MS Windows Server 2003 SP1 x64
4 GB RAM
2x iSCSI 1 Gb/sec VLANs
IOMETER
Version 2006.07.27
Access Random/Sequential
IO Size Read/Write Distribution
Specification Distribution
Exchange 2007 EDB 8 KB 90% Random 80% Read
Exchange LOG 64 KB 100% Sequential 100% Write
Exchange
Number of Servers Physical Drives RAID Protection
Simulation
Iometer 1 14-42 RAID 1+0

© 2009, Enterprise Strategy Group, Inc. All Rights Reserved.

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