Dell DL385 User Manual

TEST REPORT

MAY 2009

Initial investment payback analysis: Dell
PowerEdge R710 solution with VMware ESX vs.
HP ProLiant DL385 solution

Executive summary

Dell Inc. (Dell) commissioned Principled Technologies (PT)
to estimate how many months it would take to recapture
initial investment costs when consolidating multiple 4-year­old HP ProLiant DL385 server and storage solutions onto a
Dell™ PowerEdge™ R710 server and storage solution using
VMware ESX. In this report, we estimate both the number of
older solutions each Dell PowerEdge R710 solution can
replace and the payback period for replacing those older
solutions. We compare the following two solutions:

• Intel® Xeon® Processor X5550-based Dell
PowerEdge R710 server with 96 GB of memory
using VMware ESX and Dell™ EqualLogic™
PS6000XV storage (Dell PowerEdge R710 solution)

• AMD Opteron 254-based HP ProLiant DL385 G1
server with 4 GB of memory and HP StorageWorks
MSA30 storage (HP ProLiant DL385 solution)

KEY FINDINGS

z Each Dell PowerEdge R710 solution with

VMware® ESX™ can replace seven HP
ProLiant DL385 solutions, and could
yield a payback in under 18 months.
(See Figure 1.)

z The Dell PowerEdge R710 solution with

VMware ESX delivered seven virtual
servers, each of which yielded
performance equivalent to or better than
the performance of one physical HP
ProLiant DL385 solution. (See Figure 4.)

z The Dell PowerEdge R710 solution with

VMware ESX used little more than 1/5th
of the power, about 4/5 of the software
costs, and 1/7th of the data center rack
space of the seven HP ProLiant DL385
solutions. (See Figures 3 and 7.)

Figure 1: A single Intel Xeon Processor X5550-based Dell PowerEdge R710 solution with VMware ESX allows you to consolidate
seven AMD Opteron 254-based HP ProLiant DL385 solutions, with an initial investment payback period of under 18 months. We base
this estimation on our specific database workload.

Our test case modeled a typical enterprise datacenter with multiple legacy HP ProLiant DL385 solutions running
high-demand database workloads. The legacy servers each used 4 GB of memory. The enterprise in this test
case seeks to consolidate several of these legacy workloads onto Dell PowerEdge R710 solutions using VMware
ESX and configured with sufficient processors, memory, and storage to handle these workloads.

We used benchmark results from the Dell DVD Store Version 2.0 (DS2) performance benchmark to determine the
number of older servers with accompanying storage that a Dell PowerEdge R710 solution could replace. To
define the replacement factor, we measured the number of orders per minute (OPM) that the HP ProLiant DL385
solution could perform when running a demanding DS2 workload. We then ran DS2 in virtual machines (VMs) on
the Dell PowerEdge R710 solution and measured how many VMs this newer solution could run, while obtaining

OPM in each VM equivalent to that of the HP ProLiant DL385 solution. The total number of VMs the Dell
PowerEdge R710 solution could handle determines the replacement factor. Based on our specific workload, the
replacement factor is seven. Figure 1 depicts this replacement factor for replacing seven AMD Opteron 254­based HP ProLiant DL385 solutions with a single Intel Xeon Processor X5550-based Dell PowerEdge R710
solution, as well as the initial investment payback time of under 18 months.

Figure 2 graphs the payback period and the cost savings of the Dell PowerEdge R710 solution. Savings continue
after the initial investment payback period. By the end of year two, we project a savings of $16,268. The line
representing the Dell PowerEdge R710 solution accumulates the initial investment cost and the monthly costs of
the solution. The initial investment cost includes the list price of the server and half of the cost of the shared
storage, as well as the costs of migrating from the seven HP ProLiant DL385 solutions to the newer Dell
PowerEdge R710 solution. The line for the seven HP ProLiant DL385 solutions shows the accumulated monthly
costs of seven HP ProLiant DL385 solutions. The lines cross at the end of the payback period. Appendix C
describes the payback period calculation.

Dell PowerEdge R710 solution can pay back the initial investment in under 18

months

$140,000

$120,000

$100,000

$80,000

$60,000

$40,000

$20,000

$0

Breakeven at

$60,016

$44,557

initial investment

1 2 3 4 5 6 7 8 9 101112131415161718192021222324252627282930313233343536

Seven HP ProLiant DL385 solutio ns Dell PowerEdge R710 solution

Payback period Initial investment

2nd year savings 3rd year saving s

in 17.6 months

$16,268

savings at 2 years

$46,680

saving s at 3 years

Figure 2: The payback period and cumulative estimated costs1 for the seven HP ProLiant DL385 solutions and the Dell
PowerEdge R710 solution with VMware ESX. Lower costs and higher savings are better.

Costs for the Dell PowerEdge R710 solution include the initial investment of $44,557 plus annual costs of
$10,552, for a total year-one and -two cost of $65,660 for the solution. This solution could save approximately
$16,268 by year two over the $81,928 two-year costs for the seven HP ProLiant DL385 servers and HP
StorageWorks MSA30 storage arrays that this solution can replace. Figure 3 shows the specific areas of savings.
Specifically, it shows that the investment costs of buying the server and storage hardware and of migrating from
the older to the newer solution are offset by savings for the Dell PowerEdge R710 solution in hardware support
costs, software licenses, software support agreements, facility costs (including switch costs), energy costs, and
management costs.

1

We divide annual costs by 12 and do not adjust for inflation.

Principled Technologies, Inc.: Initial investment payback analysis: Dell PowerEdge R710
solution with VMware ESX vs. HP ProLiant DL385 solution

2

Estimated year-one and -two costs for the Dell

PowerEdge R710 solution and the HP ProLiant DL385

solutions

$90,000

$81,928

$80,000

$70,000

$60,000

$50,000

Cost

Server and sto rage hardware

Software licenses

Mig ratio n co sts

Management costs

Energy costs

Facility costs

Hardware support agreements

Sof tware support agreements

To t al

$40,000

$30,000

$20,000

$10,000

$0

Seven HP ProLiant

DL385 so lutions

$0 $43,652

$0 $2,993

$0 $905

$39,200 $5,300

$13,608 $1,894

$4,046 $450

$11,452 $2,466

$13,622 $8,000

$81,928 $65,660

Dell PowerEdge R710

solution with VMware

$65,660

ESX

Figure 3: Estimated year-one and -two costs for the seven HP ProLiant DL385 solutions and the single Dell PowerEdge R710
solution with VMware ESX. Lower costs are better.

Principled Technologies, Inc.: Initial investment payback analysis: Dell PowerEdge R710
solution with VMware ESX vs. HP ProLiant DL385 solution

3

Four benefits of the Dell PowerEdge R710 solution contribute significantly to the cost savings for this solution:

• Replaces up to seven HP ProLiant DL385 servers with HP StorageWorks MSA30 storage arrays.
The increased I/O capacity, database performance, and memory efficient design of the Dell PowerEdge
R710 solution enables consolidation of multiple older systems for the specific workload we tested.

2

• Uses slightly more than one-fifth of the power of the seven HP ProLiant DL385 solutions it
replaces. The consolidated Dell PowerEdge R710 solution provides considerable energy savings

because it requires a little more than 20 percent of the power of the older HP ProLiant DL385 solutions
that it replaces.

3

(See Figure 7.) We gain additional savings in energy costs because the test workload

requires at most half the shared storage array.

• Frees up five Windows Server® licenses. Consolidation provides software savings if the organization
maintains software assurance agreements or can reuse elsewhere the licenses that consolidation frees.
We do not include license costs for either solution, as the enterprise would already have paid for the
licenses used on the older solution and would transfer them to the new solution. We do include the costs
of ongoing software agreements. For the calculations in this paper, we assume that the target
organization maintains software agreements for OS and database software and that it can choose to
renew or cancel them at the time of consolidation.

4

• Requires approximately one-seventh of the data center rack space. The consolidated Dell
PowerEdge R710 solution requires 2u of rack space for the server and 3u for the storage array. The older
HP ProLiant DL385 solution requires 5u for each of the seven server-and-storage-array pairs, for a total
of 35u; by contrast, the single Dell R710 solution consumes only 5u.

Workload

We conducted our testing using DVD Store Version 2, an open-source application with a back-end database
component, a front-end Web application layer, and a driver layer that operates as the middle tier and actually
executes the workload. DS2 models an online DVD store. Simulated customers log in; browse movies by actor,
title, or category; and purchase movies. The workload also creates new customers. Browsing movies involves
select operations, some of which use SQL Server 2008 full-text search and some of which do not. The purchase,
login, and new customer stored procedures involve update and insert statements, as well as select statements.
The workload’s main reporting metric is orders per minute.

Because our goal was to isolate and test database server and storage, we did not use the Web application layer.
Instead, we ran the driver application on client machines directly via its command-line interface. To simulate a
heavily loaded environment, the load-generating systems ran with no think time, blasting requests as quickly as
the server could handle them. We used the default DS2 parameters and setup configuration, with the exceptions
we note in the DVD Store setup section in the Test methodology section of this report.

For client machines, we used high-end desktop systems running ESX Build 148592. Each of these client
machines ran a single instance of DS2, which spawned 32 threads. This simulated a heavily loaded environment.
Our goal was to represent the consolidation of older servers facing a demanding workload, so we adjusted the
DVD Store thread count to 12 on the older server in order to get the server to about 80 percent processor
utilization throughout the test.

We used four internal hard drives (two RAID1) on all servers, installing the OS on one volume and using the other
volume for DB logs. We configured the external storage as one large volume for the OS virtual machines and 14
25GB LUNs to be used for SQL Data volumes (2 for each of our 7 VMs).

2

The replacement factor is the number of VMs that ran on the Dell PowerEdge R710 solution, where on average each VM was equivalent to

or better than the performance of the HP ProLiant DL385 solution. The Test results section of this report details those results.

3

We multiply the replacement factor times the average power of the older server and storage solution and compare that result to the power of

a single Dell PowerEdge R710 server and Dell EqualLogic PS6000 storage array.

4

Our calculations include the costs for Windows Server 2003 R2 Enterprise Edition software assurance for the older solution and for the
upgraded Windows Server 2008 Enterprise Edition for the Dell PowerEdge R710 solution. We assume the licenses transfer between the
solutions. We base this savings claim on the seven licenses required for seven older solutions versus two licenses, each supporting four VMs,
for the Dell PowerEdge R710 server.

Principled Technologies, Inc.: Initial investment payback analysis: Dell PowerEdge R710
solution with VMware ESX vs. HP ProLiant DL385 solution

4

The DS2 driver application creates an orders-per-minute performance counter on each client machine. While the
DVD Store client application visually outputs OPM at 10-second intervals, we chose to collect this OPM metric via
a performance monitor counter on each client machine at 1-second intervals.

For more details about the DS2 tool, see http://www.delltechcenter.com/page/DVD+Store

We ran this workload on the Dell PowerEdge R710 server and the HP ProLiant DL385 server for 30 minutes.

Our test case modeled a typical enterprise datacenter with multiple legacy HP ProLiant DL385 solutions running
high-demand database workloads. The enterprise in this test case seeks to consolidate the workloads of seven of
these legacy systems onto a Dell PowerEdge R710 solution that runs VMware ESX and wants to yield a payback
in under 18 months. The enterprise seeks to balance performance and cost.

We ran the DS2 benchmark to verify that, using our test workload, the newer solution would outperform the seven
legacy systems it would replace. We tested each legacy system and each VM on the newer solution with a 10GB
database. Each VM on the newer system delivered better performance than did a single legacy system with the
same workload. We then estimated the monthly costs for maintaining the seven legacy and single new system
and estimated the payback period for the investment in the newer solution.

To model this test case, we did not compare identical solutions but instead configured the newer solution with
sufficient processors, memory, and storage to not only match or improve on the performance that the legacy
solutions deliver for the test workload but also yield a payback in under 18 months.

We selected 4GB RAM for the legacy systems as it is a realistic amount of memory for a 4 year old system. We
configured the newer solution as we expect an IT manager would when purchasing a Nehalem-based database
solution intended to replace seven HP ProLiant DL385 solutions running the high-demand databases modeled by
our workload. We selected and priced a Nehalem-based server with two processors, 96GB RAM, and modern
storage. As we expect an IT manager would, we balanced performance and cost to define the newer
configuration.

We used nearly the maximum available memory per VM that the new server could support with 96 GB (8 GB x

12) of memory, allocating 12GB of memory per VM for a total allocation of 84GB for the seven VMs. We expected
that the IT administrator in our test case would take advantage of the memory capacity of modern servers in order
to improve performance and consolidation and would select 12GB of memory per VM, enough memory that the
10GB database would be likely fit into memory,

We tested the legacy solution with HP StorageWorks MSA30 storage, which had 14 146GB 10,000 RPM disks.
We configured the newer solution with modern storage selecting a Dell™ EqualLogic PS6000XV storage array
with 15 450GB 15,000 RPM disks. The list price for that array is $55,000. We assumed that the solution we
modeled would require half that storage and include the half the costs of the storage array, including $27,500 of
the investment costs, in the payback analysis.

We tested the legacy system with 4-year old software versions and the newer system with current software
versions, assuming that the IT department in the test case would take advantage of the performance benefits of
newer software. We tested the Dell PowerEdge R710 solution running VMware vSphere 4 Advanced

5

and the HP
ProLiant DL385 solution running Windows Server 2003 R2 Enterprise Edition. We tested the Dell PowerEdge
R710 solution with Microsoft SQL Server 2008 Enterprise Edition and the HP ProLiant DL385 solution with
Microsoft SQL Server 2005. We installed Windows Server 2008 Enterprise Edition and Microsoft SQL Server
2008 in each VM on the Dell PowerEdge R710.

5

We tested with a vSphere 4.0 prior to product release using a build from the post release candidate timeframe. We anticipate that the

performance we saw with that build will be very close to that of the released product.

Principled Technologies, Inc.: Initial investment payback analysis: Dell PowerEdge R710
solution with VMware ESX vs. HP ProLiant DL385 solution

5

Our results show the advantage in performance, cost, and workload consolidation that modern hardware, OSs,
hypervisors, and database software can offer over legacy hardware and software solutions for this workload.

Test results

Microsoft SQL Server workload results

Higher results are better

8

7

6

5

4

3

2

1

0

Number of virtual machines

Server and storage solution

Figure 4: Number of virtual machines for the Dell PowerEdge R710 solution and HP ProLiant
DL385 solution. Higher numbers are better.

StorageWorks MSA30 storage solution while running the DVD Store tests. Results are from the test run that
produced the median of three OPM results. Lower CPU percentage and higher OPM numbers are better.

We tested the legacy system with a single database instance running a 10 GB DS2 database, which alone nearly
fully utilized the system’s CPU; additional instances yielded unacceptable results. To achieve maximum DS2
results on the Dell PowerEdge R710 solution, we installed one instance of SQL Server 2008 on each of seven
VMs and ran one 10GB DS2 database on each. We show the number of OPM each solution achieved during our
measurement period, minutes 24 through 29 of the test. We calculated the scores by averaging the OPM for each
SQL instance during that period. For the Dell PowerEdge R710 solution, we show the OPM results for the lowest
performing of the seven VMs. Each of the seven VMs on the Dell PowerEdge R710 solution delivered
performance equivalent to or better than the performance of the HP ProLiant DL385.

HP ProLiant DL385 with
HP StorageWorks MSA30

Dell PowerEdge R710 with
Dell EqualLogic PS6000

Figure 4 illustrates the
number of virtual machines
the Dell PowerEdge R710
solution can sustain
compared to the HP
ProLiant DL385 solution.
The Dell PowerEdge R710
solution delivered seven
virtual servers that each
produced higher OPM
results than the single
physical HP ProLiant DL385
solution.

Figure 5 shows the specific
performance results, in both
OPM and CPU percentage,
for the Dell PowerEdge
R710 server with the Dell
EqualLogic PS6000 storage
solution and the HP ProLiant
DL385 server with the HP

Performance

Run 1 Run 2 Run 3

Minimum OPM Minimum OPM Minimum OPM

HP ProLiant DL385 solution 15,864.5 14,466.8 15,849.3
Dell PowerEdge R710 solution 16,029.3 15,360.9 15,036.6

Figure 5: Minimum per database OPM performance results per server or virtual server for the Dell PowerEdge R710 solution and the
HP ProLiant DL385 solution while running the DVD Store workload. Higher OPM numbers are better.

Figure 6 shows the specific performance results, in both individual VM OPM scores and minimum OPM scores,
for the Dell PowerEdge R710 server with the Dell EqualLogic PS6000 storage array. Higher numbers are better.

Principled Technologies, Inc.: Initial investment payback analysis: Dell PowerEdge R710
solution with VMware ESX vs. HP ProLiant DL385 solution

6

Dell PowerEdge R710
solution

Run 1 Run 2 Run 3

VM 1 OPM 16,218.6 16,233.8 15,036.6
VM 2 OPM 17,261.6 17,146.0 16,496.1
VM 3 OPM 17,012.3 16,887.3 16,289.6
VM 4 OPM 17,335.5 17,040.7 16,471.6
VM 5 OPM 17,537.8 16,795.8 15,901.0
VM 6 OPM 17,093.4 17,324.7 16,325.2
VM 7 OPM 16,029.3 15,360.9 15,697.6
Minimum OPM 16,029.3 15,360.9 15,036.6

Figure 6: Individual VM scores and the minimum VM score for the Dell PowerEdge R710 solution. Higher OPM numbers are better.

Figure 7 shows the specific power results during the median run of the Dell PowerEdge R710 server with the Dell
EqualLogic PS6000 storage array and the HP ProLiant DL385 server with the HP StorageWorks MSA30 storage
array. Lower numbers are better.

To calculate the power, we average the power we recorded during our workload measurement period, minutes 24
through 29 of the test, for both server and storage, with total power the sum of both. For the idle power, we
averaged the server and storage power and then took the sum as total. For the idle power, we allowed the system
to sit for 10 minutes after booting to the desktop and then started 2 minutes of idle power capture.

Power (median results)

Server Storage Total Server Storage Total

Workload Idle

HP ProLiant DL385 solution 314 257 571 287 254 541
Dell PowerEdge R710 solution 374 496 869 223 475 698

Figure 7: Median power results, in watts, for the Dell PowerEdge R710 solution and the HP ProLiant DL385 solution while running
the DVD Store workload and while idle. Results shown are from the runs that produced the OPM scores in Figure 5. Lower numbers
are better.

Principled Technologies, Inc.: Initial investment payback analysis: Dell PowerEdge R710
solution with VMware ESX vs. HP ProLiant DL385 solution

7

Test methodology

Adjusting BIOS settings

We used the default BIOS settings on the HP ProLiant DL385 server and enabled Virtualization Technology on
the Dell PowerEdge R710 server.

Setting up and configuring the Dell PowerEdge R710 storage and Dell EqualLogic
PS6000 storage

Dell PowerEdge R710 and Dell EqualLogic PS6000 storage configuration

We installed four internal drives in the Dell PowerEdge R710 server. We attached the server to a Dell EqualLogic
PS6000 containing 16 drives. We cabled the PS6000 to a Dell PowerConnect™ 6248 switch via its four available
ports, and we cabled the server to the switch using three of the four on-board ports. We used one network port for
client data communication and one for the ESX iSCSI initiator for the OS VM drives. We then installed the Host
Integration Toolkit (HIT) on each VM for storage management and for multipath I/O support. We configured the
two remaining network ports for communication from VMs to individual database LUNs.

We configured the internal drives in two RAID 1 volumes, with two disks per volume. The PS6000 contained 16
450GB 15,000RPM drives. We configured the PS6000 in RAID 10 mode, and created 14 25GB LUNs to be used
for SQL Data volumes (2 for each of our 7 VMs) accessed via Microsoft iSCSI Initiator, which we enabled in each
VM, and one larger LUN for VM operating system virtual disks.

We installed ESX software on the first mirrored internal volume, configured the virtual disks for SQL Server
transaction logs on the second internal volume, and placed the virtual machine operating system and SQL Server
database data on the PS6000 volume.

Setting up the internal storage for operating system installation and virtual disks for SQL Server
transaction logs

1. Enter the RAID controller BIOS by pressing Ctrl+R at the prompt.

2. Highlight Controller 0, and press F2.

3. Select Create New VD.

4. Select the first two drives, select RAID level 1, tab to the OK button, and press Enter. Accept the warning
regarding initialization.

5. Select the new virtual drive, press F2, and select Initialization, Start Init.

6. Repeat steps 2 through 4 for the second internal volume.

7. Wait for the initialization operations to complete.

8. Press Escape, and choose to Save and Exit to return to the boot sequence.

Setting up the external storage

1. Using the command line, setup a storage group with the following RAID policies:

a. PS6000: RAID-10

2. Create a storage pool that you name database with the following members:

a. PS6000-1

3. Enable all network connections on the PS6000 using the following IP address scheme:

a. IP Address: 192.168.1.## (## being any number between 10 and 40)
b. Subnet Mask: 255.255.248.0

4. Create one 400GB volume in the storage pool with no snapshot reserve, and name it dvd-store

5. Create 14 25GB volumes in the storage pool with no snapshot reserve, and name each of them
dbdataVM#@ (where # is the number of the VM, and @ is the database volume – e.g., dbdata1A).

6. Enable shared access to the iSCSI target from multiple initiators on the volume.

7. Create an access control record for the volume without specifying any limitations.

8. Create a read-only SNMP community name to use for group monitoring.

Setting up the Dell PowerEdge R710 server

We installed build 148592 of a version of ESX still under development on the host server. ESX build 148592
supports the new virtualization technologies such as EPT. We also installed the vSphere client tool on a

Principled Technologies, Inc.: Initial investment payback analysis: Dell PowerEdge R710
solution with VMware ESX vs. HP ProLiant DL385 solution

8

management workstation, which was an Intel Pentium 4, 630 3.00GHz running Windows Server 2003 with 2GB
RAM.

Installing ESX Build 148592

1. Insert ESX Build 148592 DVD, and restart the computer.

2. While booting, press F11 to enter Boot Menu.

3. Press the down arrow to navigate to the appropriate boot device, and press Enter.

4. To start ESX in graphical mode, press Enter.

5. On the Welcome screen, click Next.

6. Click the checkbox to accept the terms of the license agreement, and click Next.

7. Accept the default keyboard layout, and click Next.

8. On the Custom Drivers screen, choose No to install custom drivers, and click Next.

9. You will receive a warning if you chose No. Click Yes to proceed. Drivers will load at this point.

10. Click Next.

11. Enter your serial number now, or choose to enter one later. Click Next.

12. On the Network Configuration screen, choose the NIC that you will use for system tasks. Click Next.

13. On the next Network Configuration screen, set your IP addressing information, subnet mask, DNS, and
hostname. Click Next.

14. On the Setup Type screen, choose standard setup.

15. Select the virtual disk you wish to install ESX on, and click Next.

16. On the Data Loss warning screen, click OK.

17. On the Time Zone Settings screen, choose your appropriate time zone, and click Next.

18. On the Date and Time screen, modify the date and time as you wish, and click Next.

19. Assign a root password for ESX. Optionally, add additional users if you wish, and click Next.

20. On the Summary screen, click Next to install.

21. When the installation is complete, click Next.

22. Click Finish to reboot the system.

Enabling secure shell (ssh) access for the root user

1. Log in as root to the ESX console

2. Type the following command to change to the appropriate directory:

# cd /etc/ssh

3. Edit the sshd_config file using vi. Use the following command:

# vi sshd_config

4. Press the down arrow key to move the cursor to the PermitRootLogin line, and then move the cursor to
the word no. Press the i key to insert text.

5. Type the word yes and delete the word no.

6. Press the Escape key to return to command mode.

7. Type the following command to save and exit the file:

:wq

8. Type the following command to restart the ssh service:

# service sshd restart

Installing vSphere on a management workstation

1. Configure your management workstation to be on the same subnet as your ESX server. In Internet
Explorer, navigate to the IP address of your ESX Server.

2. Accept certificate warnings, and continue.

3. Click vSphere to download the vSphere client.

4. Choose Run to install.

5. Choose your language, and click OK.

6. On the Welcome screen, click Next.

7. Accept the license terms, and click Next.

8. Enter user and organization details, and click Next.

9. Optionally, click the checkbox to install the update utility. Click Next.

10. On the Destination Folder screen, click Next.

11. Click Install.

Principled Technologies, Inc.: Initial investment payback analysis: Dell PowerEdge R710
solution with VMware ESX vs. HP ProLiant DL385 solution

9