>>
65FEBRUARY 2008 ITO AMERICA
>>
65FEBRUARY 2008 ITO AMERICA
DATACENTEROPTIMIZA-
TION
DATACENTER
OPTIMIZATION
Delivering increasing levels of
service, reducing complexity,
lowering costs, and ensuring
availability are just a few of today’s
challenges.
Data centers must be high
performance, reliable, efficient and as
scalable as services demand –without
disruption -- and easily managed from
anywhere at anytime. Finally, solutions
must satisfy adherence to regulatory
and compliance guidelines.
As many have already discovered,
there are few solutions that address
these and other related challenges
as effectively and efficiently as
modular bladed architectures. These
architectures consolidate computing
and I/O components into smaller
modular form factors. They share
common resources, such as power,
cooling and management, while
lowering costs, using less space,
and reducing overall complexity.
Generally referred to as “blades,” these
architectures can satisfy data center
requirements for desktops, servers,
networking and storage.
Countless data center professionals
have found that blades provide world-
class IT solutions. Join us as in this
series as we explore key areas where
bladed architectures provide value and
address critical data center challenges.
Learn more about how blades are
“green” as well as how blade based
technologies are changing computing.
Discover how virtualization and blades
integrate tightly together to deliver
tangible benefits and cost savings.
I.T. SOLUTIONS USING BLADES
Solving data center challenges is clearly on the minds of IT professionals globally
- Chris Hipp and Nelson Stewart
A Brief History
The blade server form factor emerged in early 2001 in the form of Transmeta powered
server blades from a start-up company in Texas called RLX Technologies. The RLX
blade server design was originally targeted at web hosting and used Transmeta
Crusoe processors, which allowed them to be eight times the density and six times
as efficient as standard rack mount servers of the day. Unfortunately for RLX,
the DotCom Bubble burst in 2001 as the company was launching its first product,
taking with it the hosting market. The company attempted to ply its web servers in
the enterprise market, but the Transmeta processors, while good for web serving,
weren’t up to the task of running back office apps.
RLX eventually changed its business plan to design
and build high performance blade servers in order to
compete head-to-head with the tier one vendors for
blade market share, and was subsequently acquired
by Hewlett-Packard in 2005. The concept of the blade
server form factor remained as a game changing and a
significant development in data center technology. The
revolution had begun.
Around the same time as the RLX server blade
introduction, Austin, TX based ClearCube Technology
launched the first PC in a blade form factor (now
referred to as the Blade PC). The ClearCube blade
PCs used standard desktop PC processors, graphics
components and a proprietary KVM-like cabling
solution to move desktop PCs out of the cubicles and
consolidated them into the data center where they
could realize the same efficiency benefits as the blade server. Today the blade PC
stands poised to launch the next major market for blade technology. The rise of the
blade PC is a topic for a future discussion.
Server Consolidation and Optimization Solutions Using Blades
Our first article explores how blade servers provide a dramatically more efficient
platform for data center server consolidations. Co-location of geographically
distributed servers into a common data center space is enabled by smaller form factor
modular blade servers delivering equivalent functionality while using significantly less
energy and data center space. Once consolidated, server blades can be optimized
through the use of virtualization software to maximize utilization of resources yielding
even greater economies of scale.
The Perpetual Challenge from I.T. Management:
“Do More, With Less”
When considering a consolidation move to blade server technologies, proper planning
and preparation goes a long way. Systems that are properly architected from the out-
set require less effort to scale later. Start by initiating an assessment of the current
“The best
candidates for
consolidation
onto blade
servers are
the apps that
can be scaled
out, not up”.
67FEBRUARY 2008 ITO AMERICA 67FEBRUARY 2008 ITO AMERICA
DATACENTEROPTIMIZA-
TION
DATACENTER
OPTIMIZATION
67FEBRUARY 2008 ITO AMERICA
overall data center
infrastructure in
order to identify
systems that have
become obsolete
and/or could be
consolidated onto newer, more efficient hardware. Replacing
older systems in your environment usually yields immediate
energy and space efficiency gains. Often, multiple servers
and their applications can be consolidated onto a single, more
powerful and efficient server. This practice involves using virtu-
alization software to optimize server blade resources. Virtual-
ization software mitigates concerns when consolidating legacy
applications and operating systems by abstracting traditional
“software to hardware” dependencies. This enables unparal-
leled platform portability and improves net consolidations. This
feature alone can easily justify initial consolidation costs, and
eliminates many headaches for IT managers.
While blade servers allow efficient
consolidation of most applications,
not all applications can or should be
consolidated onto blades. While blade
servers are available in a wide variety of
performance and expansion capabilities,
some applications may still need to be
left with large standalone or rack mount
servers, allowing them to have the
maximum of performance and peripheral expansion flexibility.
These days most, if not all enterprise data centers employ
some combination of scale up and scale out servers.
The best candidates for consolidation onto blade servers are
the apps that can be scaled out, not up. These are typically
the services that need to scale dynamically according to
changing workloads where multiple instances of the same
application are deployed in parallel. There are two basic levels
(or layers) of scaling application services when utilizing blades
and virtualization that make it the ideal solution when this
applies:
The first is the hardware. Blade servers are designed from the
ground up to be hot-swappable. The passive mid-plane design
of most blade server chassis’ allows for power, networking and
monitoring/control signals to pass through a single connector,
making adding or removing a blade server as easy as sliding it
into or out of its slot. If a server fails or additional performance
is warranted, additional blade servers are more easily added,
removed or replaced than ever before. Today, most blade
server chassis are equipped with built-in sophisticated
“First, and
foremost, blade
computing
lowers power
consumption”.
management and monitoring tools enabling instant discovery if
failures occur.
The second level is the virtualization layer. The virtualization
layer is formed when virtualization software is added on top
of blade server hardware to form a powerful solution. For
example: when the software that monitors and manages the
application service detects a failure or a surge in demand,
it can automatically and dynamically shift workloads and/
or add capacity by firing up warm spare (blades) in the
chassis. These same resources can be released back to the
“free pool” again when problem has subsided. This kind of
automated dynamic control over the hardware and software
environment can save tremendous amounts of energy, time
and money. It can also deliver a level of redundancy that is
difficult to replicate when deploying traditional servers. The
level of monitoring and control that blade servers offer is
unprecedented.
Although initial or partial blade implementations can be
more expensive, costs can be justified and recouped
over time in several ways. First, and foremost, blade
computing lowers power consumption. Since blade
servers share common chassis, power supplies,
networking and cooling, they require a lot less power than
traditional servers. Depending on the applications, when
compared with similar standalone servers, a typical blade
server can pay for itself in energy savings over the 3-year
average life expectancy of the hardware. Another benefit of
blade computing is the density that comes from the hardware’s
small form factor, resulting in more efficient use of data center
facility space that is costly to build and maintain. The gain
from space efficiency can be significant – savings go straight
to contributing to companies’ bottom line. Adding to these
the ease of deploying, monitoring and managing blades, the
technology offers a powerful ROI story to the CFOs.
As blade technology’s first industry group, the Blade Systems
Alliance is vendor-neutral and focused on expanding the
market for end-users and vendors. In 2006, Blade Systems
Alliance partnered with PPM Media and launched the industry’s
first invitational summit on blade technology in the spring of
2007. In 2008, BladeSystems Insight (BSI) will be expanded
to take place in the U.S., Europe, and Asia. The organization
is currently developing an Interoperability Lab at an end-user
site featuring Blade Systems Alliance member companies’
component solutions. The lab will be showcased at the BSI
conference in Tucson, AZ in May of 2008.
For more information see: http://bladesystems.org

Solutions_using_Blades_ITO0108

  • 1.
    >> 65FEBRUARY 2008 ITOAMERICA >> 65FEBRUARY 2008 ITO AMERICA DATACENTEROPTIMIZA- TION DATACENTER OPTIMIZATION Delivering increasing levels of service, reducing complexity, lowering costs, and ensuring availability are just a few of today’s challenges. Data centers must be high performance, reliable, efficient and as scalable as services demand –without disruption -- and easily managed from anywhere at anytime. Finally, solutions must satisfy adherence to regulatory and compliance guidelines. As many have already discovered, there are few solutions that address these and other related challenges as effectively and efficiently as modular bladed architectures. These architectures consolidate computing and I/O components into smaller modular form factors. They share common resources, such as power, cooling and management, while lowering costs, using less space, and reducing overall complexity. Generally referred to as “blades,” these architectures can satisfy data center requirements for desktops, servers, networking and storage. Countless data center professionals have found that blades provide world- class IT solutions. Join us as in this series as we explore key areas where bladed architectures provide value and address critical data center challenges. Learn more about how blades are “green” as well as how blade based technologies are changing computing. Discover how virtualization and blades integrate tightly together to deliver tangible benefits and cost savings. I.T. SOLUTIONS USING BLADES Solving data center challenges is clearly on the minds of IT professionals globally - Chris Hipp and Nelson Stewart A Brief History The blade server form factor emerged in early 2001 in the form of Transmeta powered server blades from a start-up company in Texas called RLX Technologies. The RLX blade server design was originally targeted at web hosting and used Transmeta Crusoe processors, which allowed them to be eight times the density and six times as efficient as standard rack mount servers of the day. Unfortunately for RLX, the DotCom Bubble burst in 2001 as the company was launching its first product, taking with it the hosting market. The company attempted to ply its web servers in the enterprise market, but the Transmeta processors, while good for web serving, weren’t up to the task of running back office apps. RLX eventually changed its business plan to design and build high performance blade servers in order to compete head-to-head with the tier one vendors for blade market share, and was subsequently acquired by Hewlett-Packard in 2005. The concept of the blade server form factor remained as a game changing and a significant development in data center technology. The revolution had begun. Around the same time as the RLX server blade introduction, Austin, TX based ClearCube Technology launched the first PC in a blade form factor (now referred to as the Blade PC). The ClearCube blade PCs used standard desktop PC processors, graphics components and a proprietary KVM-like cabling solution to move desktop PCs out of the cubicles and consolidated them into the data center where they could realize the same efficiency benefits as the blade server. Today the blade PC stands poised to launch the next major market for blade technology. The rise of the blade PC is a topic for a future discussion. Server Consolidation and Optimization Solutions Using Blades Our first article explores how blade servers provide a dramatically more efficient platform for data center server consolidations. Co-location of geographically distributed servers into a common data center space is enabled by smaller form factor modular blade servers delivering equivalent functionality while using significantly less energy and data center space. Once consolidated, server blades can be optimized through the use of virtualization software to maximize utilization of resources yielding even greater economies of scale. The Perpetual Challenge from I.T. Management: “Do More, With Less” When considering a consolidation move to blade server technologies, proper planning and preparation goes a long way. Systems that are properly architected from the out- set require less effort to scale later. Start by initiating an assessment of the current “The best candidates for consolidation onto blade servers are the apps that can be scaled out, not up”.
  • 2.
    67FEBRUARY 2008 ITOAMERICA 67FEBRUARY 2008 ITO AMERICA DATACENTEROPTIMIZA- TION DATACENTER OPTIMIZATION 67FEBRUARY 2008 ITO AMERICA overall data center infrastructure in order to identify systems that have become obsolete and/or could be consolidated onto newer, more efficient hardware. Replacing older systems in your environment usually yields immediate energy and space efficiency gains. Often, multiple servers and their applications can be consolidated onto a single, more powerful and efficient server. This practice involves using virtu- alization software to optimize server blade resources. Virtual- ization software mitigates concerns when consolidating legacy applications and operating systems by abstracting traditional “software to hardware” dependencies. This enables unparal- leled platform portability and improves net consolidations. This feature alone can easily justify initial consolidation costs, and eliminates many headaches for IT managers. While blade servers allow efficient consolidation of most applications, not all applications can or should be consolidated onto blades. While blade servers are available in a wide variety of performance and expansion capabilities, some applications may still need to be left with large standalone or rack mount servers, allowing them to have the maximum of performance and peripheral expansion flexibility. These days most, if not all enterprise data centers employ some combination of scale up and scale out servers. The best candidates for consolidation onto blade servers are the apps that can be scaled out, not up. These are typically the services that need to scale dynamically according to changing workloads where multiple instances of the same application are deployed in parallel. There are two basic levels (or layers) of scaling application services when utilizing blades and virtualization that make it the ideal solution when this applies: The first is the hardware. Blade servers are designed from the ground up to be hot-swappable. The passive mid-plane design of most blade server chassis’ allows for power, networking and monitoring/control signals to pass through a single connector, making adding or removing a blade server as easy as sliding it into or out of its slot. If a server fails or additional performance is warranted, additional blade servers are more easily added, removed or replaced than ever before. Today, most blade server chassis are equipped with built-in sophisticated “First, and foremost, blade computing lowers power consumption”. management and monitoring tools enabling instant discovery if failures occur. The second level is the virtualization layer. The virtualization layer is formed when virtualization software is added on top of blade server hardware to form a powerful solution. For example: when the software that monitors and manages the application service detects a failure or a surge in demand, it can automatically and dynamically shift workloads and/ or add capacity by firing up warm spare (blades) in the chassis. These same resources can be released back to the “free pool” again when problem has subsided. This kind of automated dynamic control over the hardware and software environment can save tremendous amounts of energy, time and money. It can also deliver a level of redundancy that is difficult to replicate when deploying traditional servers. The level of monitoring and control that blade servers offer is unprecedented. Although initial or partial blade implementations can be more expensive, costs can be justified and recouped over time in several ways. First, and foremost, blade computing lowers power consumption. Since blade servers share common chassis, power supplies, networking and cooling, they require a lot less power than traditional servers. Depending on the applications, when compared with similar standalone servers, a typical blade server can pay for itself in energy savings over the 3-year average life expectancy of the hardware. Another benefit of blade computing is the density that comes from the hardware’s small form factor, resulting in more efficient use of data center facility space that is costly to build and maintain. The gain from space efficiency can be significant – savings go straight to contributing to companies’ bottom line. Adding to these the ease of deploying, monitoring and managing blades, the technology offers a powerful ROI story to the CFOs. As blade technology’s first industry group, the Blade Systems Alliance is vendor-neutral and focused on expanding the market for end-users and vendors. In 2006, Blade Systems Alliance partnered with PPM Media and launched the industry’s first invitational summit on blade technology in the spring of 2007. In 2008, BladeSystems Insight (BSI) will be expanded to take place in the U.S., Europe, and Asia. The organization is currently developing an Interoperability Lab at an end-user site featuring Blade Systems Alliance member companies’ component solutions. The lab will be showcased at the BSI conference in Tucson, AZ in May of 2008. For more information see: http://bladesystems.org