A Real Comparison with Actual Machines PRIMERGY

Report
Chapter 2. Hardware Comparison
1. 2-Socket Rack-Mounted Servers
HP ProLiant DL380 G7 vs. Fujitsu PRIMERGY RX300 S6
2. 1-Socket Rack-Mounted Servers
HP ProLiant DL120 G7 vs. Fujitsu PRIMERGY RX100 S7
3. 1-Socket Tower Servers
HP ProLiant ML110 G7 vs. Fujitsu PRIMERGY TX140 S1
FUJITSU CONFIDENTIAL
0
Copyright 2012 FUJITSU LIMITED
1. 2-Socket Rack-Mounted Servers
HP ProLiant DL380 G7 vs. Fujitsu PRIMERGY RX300 S6
The PRIMERGY RX300 S6 was compared to the HP ProLiant DL380 G7 server from the
various viewpoints. The results show Fujitsu is the clear winner in the areas of cooling,
features and serviceability. Fujitsu provides hot-swap components, server cooling
efficiency and on-board failure notification LEDs, which can be good selling points.
As for hardware design, both servers are closely competitive.
When it comes to power consumption, it seems HP has an advantage over Fujitsu in being
able to provide lower-voltage servers.
FUJITSU CONFIDENTIAL
1
Copyright 2012 FUJITSU LIMITED
Tested Configurations
For display only
HP ProLiant DL380 G7
Fujitsu PRIMERGY RX300 S6
Type
Xeon E5640
Xeon E5640
Cores
8 cores (2 x 4-core)
8 cores (2 x 4-core)
6GB (3 x 2GB)
6GB (3 x 2GB)
3 x 147GB
3 x 73GB
Supported RAID levels
RAID 5
RAID 5
Number of LAN ports
(On-board LAN)
4 ports
2 ports
Net weight
21.15kg
18.95kg
-
679W
Size (mm) / Height (U)
446 x 693 x 86 / 2U
445 x 720 x 88 / 2U
List price (including tax)
778,050 Japanese yen
836,850 Japanese yen
Machine image
Processors
Memory
HDD
Max. power consumption
(Figures in data sheets)
FUJITSU CONFIDENTIAL
2
Copyright 2012 FUJITSU LIMITED
Features and Benefits of the ProLiant DL380 G7
HP DL380 G7 vs. Fujitsu RX300 S6

For display only
* Information in this page is based on data obtained from Fujitsu’s internal tests and measurements,
and Fujitsu provides no guarantee of the accuracy and integrity of the information.
Power and cooling

With a sea of sensors throughout the server and highefficient power supplies, HP produces a new level of
energy efficiency.
Optimized for space-constrained installations
G4
G7
‘A sea of sensors’
HP’s power supply that meets
or exceeds 80 PLUS GOLD
5U
2U
The latest G7 model provides 300% larger disk
capacity, yet needs less than half the space,
compared to the G4 generation.

Server management tools
“HP Systems Insight Control Suite”

Delivers best-in-class remote management performance.
Much faster remote consoles than the
older-generation products.
• 800% faster remote console
• 360% faster virtual media
An integrated solution that helps improve operational
efficiency and reduce system administration expenses.
HP Systems Insight Manager
FUJITSU CONFIDENTIAL
Remote management (HP iLO 3)
HP SmartStart
Intuitive interfaces
3
Copyright 2012 FUJITSU LIMITED
Comparison Points
For display only
HP DL380 G7 vs. Fujitsu RX300 S6
Throughout the tests, the ‘selling points’ of the HP ProLiant hardware were carefully examined against the
Fujitsu PRIMERGY system. Tested areas included the following:
■ Power consumption
• Energy-efficient power supplies.
■ Power consumption
• Power supply units with high
energy conversion efficiency.
■ Cooling
• Smart cooling: ‘A sea of sensors’
■ Cooling
• Efficient air intakes and optimized
airflow within a rack.
■ Space optimization
• Reduce space requirements.
Hardware:
ProLiant DL380 G7
FUJITSU CONFIDENTIAL
Hardware:
Fujitsu PRIMERGY RX300 S6
4
Copyright 2012 FUJITSU LIMITED
Hardware Comparisons
HP DL380 G7 vs. Fujitsu RX300 S6
* Information in this page is based on data obtained from Fujitsu’s internal tests and measurements,
and Fujitsu provides no guarantee of the accuracy and integrity of the information.
It seems Fujitsu is a clear winner on server cooling, features and ease of serviceability. On the other hand, in the
area of hardware design, HP has a slight advantage over Fujitsu.
◎ :Particularly superior
○ : Superior
△ : Equivalent
× : Inferior
Hardware comparisons
No.
Category
Tested items
×
Cooling
• Proper ventilation in servers
• Internal server temperatures
• Fan control
◎
• The HP server has a number of small ventilation holes, but the
actual temperature measurements show the Fujitsu server can
operate at much lower temperature.
• Both servers provide fan control features.
Design
• Server weight
• Noise levels
△
• The lighter-weight Fujitsu server is more space optimized.
• HP has an advantage over Fujitsu in terms of noise levels.
○
• The PRIMERGY RX300 offers hot-swap fans.
• The HP server can contain more LAN ports than Fujitsu PRIMERGY.
But LAN ports on the Fujitsu server have LEDs that can indicate link
status and the LAN speed.
• The Fujitsu and HP servers hold the same number of PCI slots.
• HP and Fujitsu are closely competitive in being able to provide
advanced security features (TPM).
◎
• The Fujitsu server ensures that notification LEDs are lit even when
the server is unplugged. This is not offered by HP.
• The dedicated fan cover on the Fujitsu server can be removed/
reinstalled while server remains in rack. On the other hand, the HP
server must be uninstalled from a rack enclosure for every
maintenance task.
• Power consumption
• Power supply efficiency
2
3
4
5
Comments
• HP has an advantage over Fujitsu in terms of server power
consumption.
• Both servers have equivalent levels of power conversion efficiency
(Just a difference of certification organizations).
Power
consumption
1
Is Fujitsu better
than HP?
Features
•
•
•
•
Serviceability
• Failure notification LEDs
• The ability to perform
maintenance tasks while the
server remains in a rack
enclosure.
FUJITSU CONFIDENTIAL
Hot swap components
Supported LAN ports
Supported PCI slots
Data security
5
Copyright 2012 FUJITSU LIMITED
Advantages of HP over Fujitsu
Server Power Consumption
HP DL380 G7 vs. Fujitsu RX300 S6
* Information in this page is based on data obtained from Fujitsu’s internal tests and measurements,
and Fujitsu provides no guarantee of the accuracy and integrity of the information. The measurement figures
represent the actual numbers obtained in this test environment.
According to actual power measured, the Fujitsu server consumed 30W more power than
the HP ProLiant DL380 (at the maximum). Power consumption of the Fujitsu server is higher
than HP’s regardless of the operational status (when running, idling or in a standby mode).
HP ProLiant DL380 G7
Max. power consumption
(Information in data sheets)
No data published
Standby
mode
Results of
actual
measurements
FUJITSU CONFIDENTIAL
Fujitsu PRIMERGY RX300 S6
7.7W
679W
<
Idle mode
97.8W
<
Intensive
workloads
260.8W
<
6
14.4W
Fujitsu-HP
difference
varies from
7 to 30 Watts.
128.2W
275.1W
Copyright 2012 FUJITSU LIMITED
Fujitsu and HP are at the same level
Power Supply Efficiency
HP DL380 G7 vs. Fujitsu RX300 S6
For display only
* Information in this page is based on data obtained from Fujitsu’s internal tests and measurements,
and Fujitsu provides no guarantee of the accuracy and integrity of the information.
The Fujitsu RX300 S6 server uses power supply units that meet the CSCI-Gold efficiency standard.
That is equivalent to 80 PLUS GOLD level that HP claims is its advantages in power efficiency.
HP offers 80 PLUS GOLD-certified
power supplies.
Power loads
Conversion rates
20%
87% or higher
50%
90% or higher
100%
87% or higher
Fujitsu uses CSCI Gold-compliant
power supply units.
=
Power loads
Conversion rates
20%
87% or higher
50%
90% or higher
100%
87% or higher
* CSCI is a global consortium dedicated to reducing the energy consumption of the IT sector started by Google and Intel in 2007.
FUJITSU CONFIDENTIAL
7
Copyright 2012 FUJITSU LIMITED
Advantages of Fujitsu over HP
Cooling (Proper Server Ventilation)
HP DL380 G7 vs. Fujitsu RX300 S6
For display only
* Information in this page is based on data obtained from Fujitsu’s internal tests and measurements,
and Fujitsu provides no guarantee of the accuracy and integrity of the information.
Due to the insufficient number of ventilation holes, the HP server fails to deliver an adequate volume
of fresh cold air required to cool the internal components. There’s no ventilation holes around the
disk area, and it will likely increase the risk of disk failure due to thermal problems.
Insufficient ventilation holes at front and
rear of the HP server. The small holes
cannot deliver sufficient t volume of fresh
cold air.
Sufficient ventilation holes are reserved at the
front and rear of the server to ensure fresh
cold air is routed to the internal components.
Front
Front
Very small vent
holes around
hard drives.
The extremely air-permeable honeycomb
design to send fresh air to hard disks.
Limited air intakes.
Rear
Rear
The rear side of the server has large
perforated ventilation holes aligned with
airflow to further improve cooling efficiency.
Small vent holes are scattered at the rear of the
server, but they are not aligned with air flow.
FUJITSU CONFIDENTIAL
8
Copyright 2012 FUJITSU LIMITED
Advantages of Fujitsu over HP
Cooling (Component Layouts)
HP DL380 G7 vs. Fujitsu RX300 S6
For display only
* Information in this page is based on data obtained from Fujitsu’s internal tests and measurements,
and Fujitsu provides no guarantee of the accuracy and integrity of the information.
Many components sit at the rear of the HP system board. The congested component layout produces
hot spots within the server chassis, consequently leading to lower mean time before failure (MTBF).
A multi-layer arrangement of
heat-generating components
will create hot spots within
the server.
Front
Front
Not many components
sit on the motherboard
to mitigate congestion.
The component layout
is optimized in align
with airflows to cool all
components efficiently
and effectively.
Rear
Many components sit at
the rear of the system
board (Near the entrance
of PCI card slots).
Rear
3
1
2
Air-intake holes are
arranged between
PCI slots, allowing
fresh cold air blowing
to PCI cards.
6
4
5
PCI slots
FUJITSU CONFIDENTIAL
A few ventilation
holes near the
congested area.
7 6 5 4 3 2 1
PCI slots
9
Copyright 2012 FUJITSU LIMITED
Advantages of Fujitsu over HP
Cooling (Airflow Efficiency)
HP DL380 G7 vs. Fujitsu RX300 S6
For display only
* Information in this page is based on data obtained from Fujitsu’s internal tests and measurements,
and Fujitsu provides no guarantee of the accuracy and integrity of the information.
Thermal design of the HP server is not optimized (Fewer ventilation holes than Fujitsu PRIMERGY,
non-straight airflow paths). That makes it difficult to deliver air volume required to cool all internal
components, and produces hot spots within the server chassis.
The multi-layer arrangement of components will
produce hot spots near the entrance of PCI slots.
Besides, cold air is not delivered directly to this
area due to the internal server design,
consequently PCI cards may have lower MTBF.
CPU
Fresh air flows
between CPUs.
FAN
Hot air
discharge
Cold air intake
PCI
PCI
CPU
Hot air discharge
Hot
spots
Cold air intake
No space
between
CPUs.
PSU
Hot air
discharge
Hot air discharge
Not many components sit on the motherboard
to avoid a congested layout. The Fujitsu server
uses a straight-line cooling method to streamline
cold air movements within the server chassis.
FAN
PSU
Hot air discharge
Hot air generated from
the server circulates
within a server rack.
FUJITSU CONFIDENTIAL
10
Copyright 2012 FUJITSU LIMITED
Advantages of Fujitsu over HP
Cooling (Internal Server Temperatures)
HP DL380 G7 vs. Fujitsu RX300 S6
For display only
* Information in this page is based on data obtained from Fujitsu’s internal tests and measurements,
and Fujitsu provides no guarantee of the accuracy and integrity of the information. The measurement figures
represent the actual numbers obtained in this test environment.
Server temperature measurements revealed that the HP server operates at a 10˚C higher temperature
(Max.) than the Fujitsu server. During the test, both servers with their default settings were identically
loaded. A scientific study (Arrhenius Theory of Reaction Rate*) says lifetime of temperaturedependent components (i.e. Capacitors) is halved for every 10 degree Celsius increase in temperature.
High
Failure
Temperature
Fujitsu PRIMERGY is much
less susceptible to thermal
issues, allowing longer
component life.
The HP ProLiant server
operates at 10˚C higher (Max.)
than the PRIMEGY server.
Low
Short
Component life lengths
Long
* Arrhenius theory deals with the temperature dependence of the reaction rate constant, which can be used for accelerated aging test.
FUJITSU CONFIDENTIAL
11
Copyright 2012 FUJITSU LIMITED
Advantages of Fujitsu over HP
Design (Server Weight)
HP DL380 G7 vs. Fujitsu RX300 S6
For display only
* Information in this page is based on data obtained from Fujitsu’s internal tests and measurements,
and Fujitsu provides no guarantee of the accuracy and integrity of the information. The measurement figures
represent the actual numbers obtained in this test environment.
When comparing two identically rack-mounted solutions (15 Fujitsu/HP servers in a 19-inch rack),
the Fujitsu rack system was 44kg lighter than the HP system.
Fujitsu servers seem to be more suitable for customers with floor space constraints.
* Typical floor load limits in normal office buildings:
300 to 500kg per square meter.
15 units x HP ProLiant
DL580 (25.2kg-per-unit) +
1 x 40U rack (150kg)
= 528kg
This configuration cannot
be installed in an office
space with floor loading
capacity of 500kg/m².
>
15 units x PRIMERGY RX300 S6
(22.3kg-per-unit) +
1 x 40U rack (150kg)
= 484.5kg
The Fujitsu servers are lighter
than HP’s, allowing higherdensity server deployments.
This configuration can
be installed in an office
space with floor loading
capacity of 500kg/m².
* These calculations assume the rack enclosure weighs 150kg
net, and installs in 1m².
* Total weights shown above include rack rails and cable arms.
FUJITSU CONFIDENTIAL
12
Copyright 2012 FUJITSU LIMITED
Advantages of HP over Fujitsu
Design (Server Acoustic Noise Emissions)
HP DL380 G7 vs. Fujitsu RX300 S6
For display only
* Information in this page is based on data obtained from Fujitsu’s internal tests and measurements,
and Fujitsu provides no guarantee of the accuracy and integrity of the information. The measurement figures
represent the actual numbers obtained in this test environment.
Despite the Fujitsu server generating slightly higher noise levels than the HP server, it is still
at an acceptable level for deployment in normal rack installations.
Noise comparisons: Areas around server chassis measured in decibels (dBA).
HP ProLiant DL380 G7
Fujitsu PRIMERGY RX300 S6
Idle mode (dBA)
33.27
Full operation (dBA)
35.49
<
<
47.42
47.59
* These figures are for reference only.
Equivalent sound levels (For reference)
30dBA
40dBA
50dBA
60dBA
Suburbs after midnight
Library
Quiet office space
Quiet car interior
FUJITSU CONFIDENTIAL
13
Copyright 2012 FUJITSU LIMITED
Advantages of Fujitsu over HP
Features (Supported LAN Ports)
HP DL380 G7 vs. Fujitsu RX300 S6
For display only
* Information in this page is based on data obtained from Fujitsu’s internal tests and measurements,
and Fujitsu provides no guarantee of the accuracy and integrity of the information.
With LAN ports on the HP server, only link status can be viewed. But the LAN ports on Fujitsu
PRIMERGY have LEDs that indicate data transmission speed.
1Gbps
?
Rear
Rear
This LED indicates the speed
of the currently running LAN
in two distinctive colors.
100Mbps
1Gbps
Only link status
can be viewed.
Link status
To check LAN speed, it may be
required to run some sort of speed
test or a bandwidth meter within
the operating system.
FUJITSU CONFIDENTIAL
Data transmission
speed
Users can check the LAN port link status and
connection speed at a glance.
14
Copyright 2012 FUJITSU LIMITED
Advantages of Fujitsu over HP
Serviceability (Failure Notification LEDs)
HP DL380 G7 vs. Fujitsu RX300 S6
For display only
* Information in this page is based on data obtained from Fujitsu’s internal tests and measurements,
and Fujitsu provides no guarantee of the accuracy and integrity of the information.
The HP system board doesn’t have failure notification LEDs to notify a faulty component, and thus the
wrong component could be replaced by mistake. The system board on Fujitsu PRIMERGY has groups of
notification LEDs that make it easy, quick and accurate to identify problems.
The front panel
displays health
status of
components.
HP System Insight
Display at the front
of the server
System Insight Display to
indicate overall system health.
A label on the underside
of the server cover
provides information on
LED location.
??
Only disk
failures can
be notified.
System
Board
FUJITSU CONFIDENTIAL
15
7 groups of
components have
notification LEDs.
Underside of the server cover
Copyright 2012 FUJITSU LIMITED
Advantages of Fujitsu over HP
Serviceability (System Board Maintenance)
HP DL380 G7 vs. Fujitsu RX300 S6
For display only
* Information in this page is based on data obtained from Fujitsu’s internal tests and measurements,
and Fujitsu provides no guarantee of the accuracy and integrity of the information.
To replace the system board assembly of the HP server, it must be removed completely from a
rack enclosure. Service technicians need to prepare adequate flat space when conducting tasks,
and re-install the removed assembly into the rack after the maintenance.
Maintenance tasks can be completed while
the server remains in a rack enclosure.
To reach the metal latch at the rear of
the server, it is necessary to pull out the
server tray completely from a rack.
Need a driver to
release the latch.
Simply press & release the lock
at the center of the server
(Drivers are not required).
Small handles make
it difficult to pull out
the server tray.
The ‘rightsized’ fingerpull handles
56cm
The server-front
instructions assist
users to conduct
replacement tasks.
Front
FUJITSU CONFIDENTIAL
16
Copyright 2012 FUJITSU LIMITED
Recap (HP DL380 G7 vs. Fujitsu RX300 S6)
* Information in this page is based on data obtained from Fujitsu’s internal tests and measurements,
and Fujitsu provides no guarantee of the accuracy and integrity of the information.
The following summarizes the competitive advantages of the Fujitsu PRIMERGY RX300 S6 system.
To differentiate your proposal in competition against HP ProLiant DL380 G7, high reliability and
excellent serviceability of the PRIMERGY server can be good selling points.
 Server power consumption
 Equivalent levels:
Specifications of power conversion efficiency.
 HP’s advantages:
Lower power consumption with lower noise levels.
 Cooling
 Fujitsu’s advantages: Efficient airflow inside a rack; Server operation at lower temperatures.
 Design
 Fujitsu’s advantages: Lighter weight.
 HP’s advantages:
Quieter server operations.
 Features
 Fujitsu’s advantages: Standard LAN ports that are equipped with LEDs to indicate LAN connection
speed.
 Serviceability
 Fujitsu’s advantages: Failure notification LEDs on the Fujitsu server to assist quick, easy and
accurate problem detection & identification; Support for hot-swap
components.
FUJITSU CONFIDENTIAL
17
Copyright 2012 FUJITSU LIMITED

similar documents