Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Executive Overview
Contents
Executive Summary 3
Introduction to Hitachi Compute Blade 500 4
System Overview 4
Enterprise-Class Capabilities 5
Data Center Applications 6
Balanced, Modular Architecture 6
Chassis Design 7
Standard Server Blades 8
CB520X SMP Blade 8
Storage Expansion Blade 9
PCIe Expansion Blade 10
Flexible I/O 10
Mezzanine Cards 10
L3 Gigabit Ethernet Switch 11
10GbE Data Center Bridging (DCB) Switch Module 12
Eco-Efficient Power and Cooling 12
Hitachi Compute Blade Management 12
HTML-Based Web Interface 12
Logical Partitioning Embedded Virtualization Technology 13
Overview and Key Benefits 13
Performance and Security Features 14
Hitachi Compute Systems Manager 15
Unified Dashboard 15
Deployment Manager 16
N+1 or N+M Cold Standby Failover 16
Active Power Management 17
Hitachi Compute Blade 500 Specifications 18
For More Information 19
WHITE PAPER 3
Executive Summary
Hitachi Compute Blade 500 (CB 500) helps you manage rising costs in your data center. It does this by joining the
latest technology features with the innovative design, rock-solid engineering and high manufacturing quality for which
Hitachi is famous. CB 500 brings new levels of energy efficiency and reliability together with unique features, such as
Hitachi logical partitioning (LPAR) hardware-embedded virtualization technology. It lets you consolidate systems with
different I/O configurations and provides a balanced system architecture that pushes performance to new extremes.
Hitachi Compute Blade presents a unique combination of built-in virtualization, massive I/O bandwidth, large memory
capacity, browser-based point-and-click management, and unprecedented configuration flexibility. It allows you to
extend the benefits of blade computing to new areas of the enterprise data center, including mission-critical applica-
tion and database servers.
CB 500 seamlessly integrates network, power and server resources into a single, space-efficient, flexible solution,
and it also offers the very latest in power-saving features and capabilities. With sophisticated, built-in reliability, avail-
ability and serviceability (RAS) features, it reduces the risk of unplanned downtime for mission-critical applications.
CB 500 is one of the most flexible systems on the market today.
This white paper provides a high-level overview of the key features and capabilities that make Hitachi Compute Blade
500 a preferred choice over any other blade or rack server platform on the market today. And it shows how CB 500
delivers new opportunities for companies of all types and sizes.
WHITE PAPER 4
System Overview
Hitachi Compute Blade 500 with logical partitioning (LPAR) combines all the benefits of virtualization with all the
advantages of the blade server format: simplicity, flexibility, high compute density and power efficiency. This combi-
nation allows organizations to consolidate more resources, extend the benefits of virtualization solutions (VMware,
Microsoft Hyper-V and so forth) to more areas of the enterprise data center, and cut costs without sacrificing
performance.
Figure 1. Powered by Intel Xeon E5 and E7 series processors, Hitachi Compute Blade 500 packs up to 352 cores into a
fully loaded system.
By powering the system's blade modules with Intel Xeon E5 and E7 series processors and radically increasing the I/O
performance and memory capacity, Hitachi has made it possible to use virtualization to consolidate application and
database servers. These are areas where effective consolidation was difficult in the past. By removing performance
and I/O bottlenecks, CB 500 with logical partitioning reduces the administrative burden in the data center and opens
up new opportunities for increasing efficiency and utilization rates.
Optional mezzanine I/O cards connect to the shared midplane of CB 500 to allow each blade to access common
redundant shared network switches. This capability simplifies connection, saves money and eases administration. To
enable I/O versatility, CB 500 supports two pairs of redundant bays for internal switches, which can support IP net-
working (1Gb or 10Gb Ethernet), Fibre Channel or converged fabric I/O.
Built-in logical partitioning is a key feature of Hitachi Compute Blade 500. Hitachi logical partitioning is the industry's
first embedded hardware-based and hardware-assisted virtualization solution, based on proven, mainframe-class
technology. There is no complex installation process: It just runs, right out of the box. Unlike emulation solutions, it
WHITE PAPER 5
doesn't slow application performance down. And because no software hypervisor is required, your operating system
(OS) and applications don't require special drivers or software. The identical software stack is used in a logical parti-
tion as on a blade without logical partitioning enabled, simplifying management, and reducing costs of operation and
administration.
Logical partitioning complements existing software-based virtualization solutions and can be used in conjunction with
VMware, Hyper-V or Red Hat KVM. You can even combine multiple virtualization platforms in a single chassis, provid-
ing the greatest virtualization flexibility in the industry.
CB 500 also offers the very latest in power-saving features and capabilities. With industry-leading 94% efficiency, the
power supplies have been awarded the 80 PLUS Platinum certificate and conform to the Climate Savers Computing
Initiative Gold standard.
CB 500 is not only powerful and efficient, but manageable and flexible, too. You can configure and administer CB 500
via a Web-based HTML browser that supports secure encrypted communications, or leverage the optional Hitachi
Compute Blade management suite to manage multiple chassis via a unified graphical user interface (GUI) based
interface.
You can deploy CB 500 at the edge, application or database tiers, or all three at the same time. You don't need to
choose between Microsoft Windows or Linux: You can run both in the same chassis. And CB 500 integrates effort-
lessly into your network I/O infrastructure thanks to its adaptable and highly expandable I/O architecture.
Enterprise-Class Capabilities
Hitachi Compute Blade 500 is a true enterprise-class blade server, and it is important to understand exactly what that
term means. It is "enterprise-class" in terms of performance, scalability, reliability and configuration flexibility, as out-
lined below:
Performance. CB 500 supports blades based on the latest and most powerful Intel Xeon E5 and E7 series pro-
cessors with up to eight CPUs per blade (in SMP mode). It meets the performance needs of large-scale systems
that require extremely high compute power and I/O today. The extensible CB 500 architecture can support multiple
blade types, including future generations of Intel processor. Standard-width CB 500 blades can also be expanded
to support additional high-density disk (HDD) storage or additional PCI slots with an expansion blade installed in an
adjacent blade slot.
Scalability. The robust, rack-mountable 6U chassis of CB 500 houses up to eight server blade modules; each
standard-width blade may have up to two sockets, with up to 22 cores per socket. Double-width blades can
support up to two E7 processors with up to 15 cores per socket. Each blade may be configured to support up to
30 logical partitions. Memory is expandable up to 48 DIMMs in the CB520X blade, allowing up to 1,536GB to be
configured per blade using 32GB DIMMs. Each standard-width blade supports up to two I/O mezzanine cards to
connect to the chassis (four in the double-width blade), allowing each standard blade to support up to four I/O-
switch modules (usually configured as two redundant pairs).
Reliability. CB 500 chassis is fully redundant and components are hot-swappable. These components include:
redundant switch and management modules, extremely reliable backplane and I/O, N +1 or fully redundant power
supply modules, and N+M blade failover protection (referring to "M" backup blades for every "N" active server
blades, so failover is cascading). In the event of blade hardware failure, the system automatically detects the fault
and identifies the problem by indicating the faulty module, allowing immediate failure recovery. In addition, the
CB520Hv3 blade takes advantage of the latest high availability features built into the Intel Xeon processor, such as
memory mirroring and rank sparing.
Configuration flexibility. CB 500 supports Windows and/or Linux OSs, and a wide range of virtualization solu-
tions, including native LPAR, providing a high level of flexibility and investment protection. The system can easily be
configured to the exact number of sockets, processor cores, I/O slots, memory and other components required
WHITE PAPER 6
to optimally support your application without bottlenecks. The chassis can be configured and managed via simple
GUI HTML-based Web interface, which is seamlessly integrated with the Hitachi Command Suite tools used to
manage Hitachi storage products.
Workload flexibility. The enterprise-class capabilities of CB 500 make it suitable for a wide variety of applications.
CB 500 allows you to run your most demanding workloads, whether they are I/O-intensive applications such as
online transaction processing and Web serving, or high-performance computing (HPC), with extremely high per-
formance, reliability, manageability, scalability and flexibility. Because of this extreme flexibility, CB 500 is the ideal
platform to run mission-critical applications and consolidate systems at the edge, application or database tiers ...
or all three.
Figure 2. All Hitachi Compute Blade 500 modules are redundant and hot-swappable, providing high reliability and uptime.
Chassis Design
With a 6U chassis that is 19-inch rack compatible, Hitachi Compute Blade 500 accommodates up to eight server
modules or blades, which are populated with the latest Intel Xeon E5 series processors (see Figure 3). Each
standard-width Intel Xeon E5 series processor-based module supports up to two processors and up to 18 cores per
socket, for a capacity of up to 288 cores per chassis.
Figure 3. Hitachi Compute Blade 500 chassis layout allows for 19-inch rack compatibility: front (left) and rear (right).
WHITE PAPER 8
Figure 4. Hitachi Compute Blade 500 employs the CB520H server module.
The CB520X blades are full width and each contains two Intel Xeon E7 series processors with 48 DIMM slots, two
HDD bays and up to four optional mezzanine slots.
The CB520X blades feature a front-side connector, which can be used to connect either two or four blades to create
a single four- or eight-socket blade system (see Figure 6).
WHITE PAPER 9
Figure 6. Hitachi Compute Blade CB520X allows incremental expansion from two- to eight-socket blade configurations.
With support for the Microsoft Windows OS and Red Hat Linux, Hitachi Compute Blade 500 gives you the option
of running two of the most popular OSs at the same time and in the same chassis for multiple applications. Or, you
can run them (using Hitachi LPAR) on the same blade. For example, in a virtualized environment on a single active
server module, you can allocate 50% of CPU resources to Windows and 50% to Linux applications. And you can
change the allocation dynamically as workload requirements shift. This provides flexibility for accommodating spikes
in demand for specific application services. In addition, the system can be configured to dynamically reassign unused
CPU resources from one partition to other partitions where CPU resources are under stress, and back again on
demand.
Figure 7. Hitachi Compute Blade server's CB520H blade capability may be extended with a storage expansion blade.
WHITE PAPER 10
The storage expansion blade occupies the slot adjacent to its host compute blade, effectively creating a combined
double-wide blade. Up to six 900GB SAS drives may be accommodated, in addition to the two drive bays built into
the CB520H compute blades.
Figure 8. Hitachi Compute Blade server's CB520H blade capability may be extended with a PCIe expansion blade.
The PCIe expansion blade occupies the slot adjacent to its host compute blade, effectively creating a combined
double-wide blade.
Flexible I/O
Hitachi Compute Blade 500 server modules have the ability to support a variety of switch modules and configura-
tions. Each blade can access a common, unified, shared-switch network across all the blades while also supporting
unique I/O interfaces. This capability allows you to simplify the switch interface management and physical wiring. I/O
switch options include IP networking, Fibre Channel and converged data center bridged networking. These simplify
the data center wiring further and allow SAN and NAS to share the same network. This capability is particularly useful
if you have separate server, storage and network organizations, allowing storage resources to be managed outside of
the box.
Mezzanine Cards
For additional I/O versatility, Hitachi Compute Blade 500 provides two slots for optional mezzanine I/O expansion
cards (see Figure 9). The server allows plenty of bandwidth for the most demanding I/O-intensive applications, such
as online transaction processing, high-performance computing, simulations and modeling, or batch jobs with heavy
transactional loads. (When both mezzanine slots are occupied, the onboard LAN is disabled.)
WHITE PAPER 11
Hitachi Compute Blade embedded gigabit Ethernet switch can be configured for high availability and fault tolerance
when a second redundant switch module is added. A single switch interconnects one (of two) gigabit Ethernet con-
nection from each blade server module (up to eight total).
WHITE PAPER 12
Dual-speed (1Gb/sec or 10Gb/sec) capable external ports are well suited for data centers that are migrating to next-
generation 10Gb/sec high-performance architectures and are supported by both 1Gb/sec RJ-45 connectors and
10Gb/sec twinax copper or optical connectors for ultimate flexibility.
Data center bridging and multihop Fibre Channel over Ethernet (FCoE) capabilities enable lossless unified storage
connectivity, and convergence of storage and LAN traffic to reduce connectivity costs. Its automatic migration of port
profiles (AMPP) feature simplifies virtualized server management by enabling seamless virtual machine (VM) mobility,
further simplifying management and reducing operational costs.
Naturally, the DCB switch is integrated with Hitachi chassis management software, enabling end-to-end management
of CB 500.
CB 500 also features the power-capping function, which optimizes the power consumption from server blades by
controlling the frequency (CPU steps) according to processor service rate. This function saves approximately 16% of
power consumption by server blades.
In addition, CB 500 can provide real-time monitoring and historical data of power consumption in the server chas-
sis. The visualized information helps reduce equipment costs, enabling facility planning for equipment installation and
expansion.
solutions with Hitachi storage (see Figure 10). This graphical interface simplifies system configuration and supports
secure and encrypted communications, ensuring that only authorized users can access resources and only autho-
rized resources can be accessed.
Figure 10. Hitachi Compute Systems Manager provides an HTML-based Web interface.
KVM (keyboard-video-mouse) remote management of the blades is supported via IP with either a client-based inter-
face, which can be installed onto a remote computer, or via the Java-based Web interface, which is accessible from
any standard Web browser. The tools also support power control as well as virtual media capabilities, allowing local
drives to be connected directly to the server.
Logical partitioning delivers high performance, extreme reliability, security and transparent virtualization for the Intel
Xeon processor E5 series server modules on Hitachi Compute Blade 500. It supports numerous guest OSs, including
WHITE PAPER 14
Red Hat Linux and Windows, and it does not intrude at the OS level, thus maximizing security, performance and
simplicity. It keeps the data center agile and responsive to fast-changing business requirements. Whether the environ-
ment is running on a bare-metal server or on LPAR, there is no need for additional drivers or changes to applications.
Because logical partitioning is built into the blade firmware, there is no installation and setup required, reducing total
acquisition costs and simplifying the move to a partitioned environment. You can activate logical partitioning or de-
activate it quickly and easily via the management console.
Logical partitioning allows you to create multiple independent logical partitions (LPARs), which run on specific hard-
ware resources uniquely allocated to that partition by the administrator. Each logical partition appears to its operating
system to be an independent physical server.
CB 500 offers two deployment models for logical partitioning: The "Essential" model is included with each system
and allows you to create two LPARs per blade at no additional cost. The "Enterprise" model requires an additional
license and allows the creation of up to 30 LPARs per blade on CB 500.
Logical partitioning is firmware based and thus has a natural performance advantage over host-emulation virtualiza-
tion offerings, since guest OSs can be directly executed without host intervention. Logical partitioning leverages Intel
Virtualization Technology (VT) to ensure that processor performance is optimized for the virtual environment.
Equally important, this technology utilizes a standard filesystem (NTFS, the standard filesystem of modern Microsoft
Windows OSs) and standard OS drivers, simplifying the administrator's job significantly. With logical partitioning there
is no need for exclusive and costly file management, migration and/or backup tools for virtualization platforms.
Logical partitioning is fast to install, configure and deploy since all virtualization files are stored in the firmware and
there is no need to install or upgrade each blade with virtualization software. It also means better performance
because less overhead is required for filesystem format conversions or driver emulations to consume precious
CPU cycles. Drivers required are the same for either standard or partitioned environments, adding to management
simplicity.
This technology can partition physical server resources by constructing multiple LPARs, which are physically sepa-
rated, with each environment running independently. A different operating system can run on each logical partition
within a single physical server.
Logical partitioning also inherits the concept of "direct execution" from the mainframe world. This means that user-
level application code and OS requests are directly executed on the processor, increasing performance. Moreover,
Hitachi LPAR technology has been tuned specifically for Hitachi Compute Blade 500 and has been extensively tested
in enterprise production environments.
To increase security, each logical partition has dedicated physical memory and cannot access the memory of other
partitions. This ensures that no unauthorized access to a given device or memory is possible. Files are stored in their
native formats and assigned to a separate storage logical unit (LUN). Each can be assigned with different security
settings. This is a more secure approach than that used in some solutions, which employ a single file to contain all
virtual machine data.
Unlike software-based virtualization solutions, a separate management interface for LPAR is built into the hardware.
This interface resides on a different physical network than OS traffic, ensuring that user data is completely and iso-
lated from any management traffic, and adds yet another dimension of protection and security.
WHITE PAPER 15
Unified Dashboard
HCSM allows the various CB 500 system components to be managed through a unified interface, which is seam-
lessly integrated with Hitachi Command Suite (see Figure 11). When rack management is used, an overview of all
Hitachi Compute Blade racks, including which servers, storage and network devices are installed, can be quickly and
easily obtained. In the event of any system malfunction, the faulty part can be located at a glance.
In addition, HCSM software provides the ability to define and manage the logical system configuration of each ele-
ment to be managed by using the service name. With traditional blade servers, management of both the logical
system and the system's physical resources is required. When definitions are made with service names (such as
sales or stock) within Hitachi Compute Blade management suite, there is no longer any need for administrators to
concern themselves with the management of physical resources.
HCSM provides centralized system management and control of all server, network and storage resources. This
includes the ability to set up and configure servers, monitor server resources, integrate with enterprise management
software (SNMP), phone home and manage server assets.
Figure 11. Hitachi Compute Systems Manager offers a unified view of platform management.
Deployment Manager
Deployment manager allows the mass deployment of system images, patches and hot-fix updates for fast, effective
server installation and patch management. You can deploy system images and updates across multiple chassis in
multiple locations. In addition, Hitachi also offers backup capabilities with the deployment feature.
With the N+M cold standby function, there are "M" backup server blades for every "N" active server blade, so failover
recovery is cascading. In the event of multiple hardware failures, the system automatically detects the fault and identi-
fies the problem by indicating the faulty server blade, allowing immediate failure recovery. This approach can reduce
total downtime by enabling the application workload to be shared among the working servers.
WHITE PAPER 17
This feature allows a power consumption ceiling to be specified. The ceiling may be lower than the theoretical
maximum power consumption of a maximally configured chassis, which would normally dictate power and cooling
provisioning. Data center designers can thus safely provision to a lower power and cooling load and benefit from the
ensuing cost savings.
The Hitachi power capacity expansion feature removes this limit and allows the CB 500 system to be safely con-
figured to its theoretical maximum limit (see Figure 12). The configuration includes power supply redundancy and
delivers the maximum real density available.
Maximum real density is achieved by responding to a failure event within milliseconds to limit chassis power
consumption to the remaining available power. In normal use, power from the reserve supplies allows maximum
performance of all configured blades.
Figure 12. Power capacity expansion feature allows maximum density blade configurations to be supported with
redundant power.
WHITE PAPER 18
Chassis
Size 6U (rack mountable)
Dimensions (W x D x H) 447mm x 820mm x 266mm
Operating Temperature 5C ~ 40C
Humidity 20% - 80%
Server Blade Modules Up to 8 server blade modules
Management Modules 1 standard, 2 maximum (redundant)
Management LAN Interface 2 Ethernet ports
Maintenance LAN Interface 1 Ethernet port
Maintenance Serial Interface 1 serial interface
Cooling Fans 6 standard
Switch Modules 2 standard, 4 maximum
Hitachi 1Gb LAN Switch Module
Internal Ports 16 ports
External Ports 4 ports
Hitachi 1Gb or 10Gb LAN Switch Module
Internal Ports 16 ports (1Gb)
External Ports Four 1GbE ports
Two 10GbE ports
Hitachi Converged Fabric Switch Module Up to 4 modules
Internal Ports 16 ports
External Ports 8 ports
Power Supplies Up to 4 power supply modules
(N+1 fully redundant)
80 PLUS Platinum efficiency rating
Input Voltage 200 - 240VAC, single phase
Frequency 50 - 60Hz
WHITE PAPER 19
Item
HITACHI is a trademark or registered trademark of Hitachi, Ltd. Microsoft, Windows, Windows Server and Hyper-V are trademarks or registered trademarks of Microsoft Corporation. All
other trademarks, service marks, and company names are properties of their respective owners.
WP-442-J DG April 2016