"SEEDING CLOUDS ON POWER SYSTEMS WITH IBM SMARTCLOUD™ ENTRY"
1. WHITE PAPER
SEEDING CLOUDS ON POWER SYSTEMS
WITH IBM SMARTCLOUD™ ENTRY
custom consulting services
January 2012
PREPARED FOR Executive Summary
Many organizations are genuinely excited about the potential for private clouds, anticipating
IBM greater responsiveness for end users and reduced costs for administrators. However, they are
also concerned with the risk of taking on added complexity as they make the transition to cloud
computing, which by definition introduces new management processes that fundamentally
TABLE OF CONTENTS redefine how resources are assigned to workloads.
Executive Summary.............................1 IBM recognizes that many users are not ready to skip immediately to the deployment of
advanced private clouds, despite their potential benefits. More likely, depending on their culture
Cloud Cover in the Enterprise.............1 and priorities, organizations will begin to adopt cloud computing with simpler initiatives, and
progress gradually to more sophisticated cloud deployments. IBM has therefore introduced an
Virtualizing Critical Workloads ...........2
offering called IBM SmartCloud™ Entry, a scaled-down software package that enables
Scalable Virtualization...........................3 organizations to deploy clouds rapidly and with minimal disruption to existing infrastructure and
Reliable Virtualization............................3 administrative processes.
Virtual Infrastructure ...........................4 IBM SmartCloud Entry on Power Systems™ servers leverages the proven capabilities of
PowerVM® virtualization software to enable quick and easy deployment of private clouds.
Making the Leap from
SmartCloud Entry promises a relatively low-risk method of introducing the key operational
Virtualization to Cloud Computing......4
processes required for cloud computing, such as standardizing IT services, building libraries of
Virtualizing Critical Workloads these services, and implementing procedures for users to provision these services on their own.
with IBM PowerVM...............................5 By using the superior scalability and reliability of Power Systems as a foundation for their initial
cloud deployments, users can plan sufficient headroom in order to achieve transformational
Virtualization Management economies of scale in their IT infrastructure. Users can thus have confidence that they will
on IBM Power Systems........................6 reach these economies of scale as they pursue more and more advanced cloud computing
approaches.
Seeding Clouds on
IBM Power Systems.............................7
Cloud Cover in the Enterprise
The IDEAS Bottom Line .......................9 Cloud computing is entering its next phase of maturity, as organizations start to develop
strategies for deploying production workloads on cloud-based infrastructures. While users have
a long-term interest in tapping into public cloud services, many organizations are still focusing
primarily on the deployment of private clouds, in which server, storage, and network resources
are consolidated into a single pool that workloads can draw upon as needed. Eventually, as
private clouds mature, they could evolve into “hybrid” clouds that have the ability to send some
workloads to public clouds where appropriate, ideally with the same frameworks and controls
used to allocate on-premise resources. In the meantime, a key benefit of private clouds is their
ability to enable some degree of self-service, while also accounting more precisely for the
resources consumed by users and workloads. By eliminating much of the administrative
overhead required to deploy services, and more precisely allocating the resources needed to
support these services, private clouds can introduce significant cost savings and improve an
organization’s responsiveness to changing business conditions.
This document is copyrighted by Ideas International, Inc. (IDEAS) and is protected by U.S. and international copyright laws and conventions. This document may not be copied, reproduced, stored in a retrieval system, transmitted in
any form, posted on a public or private website or bulletin board, or sublicensed to a third party without the written consent of IDEAS. No copyright may be obscured or removed from the paper. All trademarks and registered marks of
products and companies referred to in this paper are protected.
This document was developed on the basis of information and sources believed to be reliable. This document is to be used “as is.” IDEAS makes no guarantees or representations regarding, and shall have no liability for the accuracy
of, data, subject matter, quality, or timeliness of the content. The data contained in this document are subject to change. IDEAS accepts no responsibility to inform the reader of changes in the data. In addition, IDEAS may change its
view of the products, services, and companies described in this document.
IDEAS accepts no responsibility for decisions made on the basis of information contained herein, nor from the reader’s attempts to duplicate performance results or other outcomes. Nor can the paper be used to predict future values or
performance levels. This document may not be used to create an endorsement for products and services discussed in the paper or for other products and services offered by the vendors discussed.
2. SEEDING CLOUDS ON POWER SYSTEMS WITH IBM SMARTCLOUD™ ENTRY JANUARY 2012
Cloud computing is often associated with virtualization. Most definitions of cloud computing
By eliminating much of the also cover approaches such as software as a service (SaaS) and platform as a service (PaaS),
which do not necessarily depend on virtualization. However, for one definition of cloud
administrative overhead computing, infrastructure as a service (IaaS), virtualization provides a powerful means of
required to deploy services, adapting existing infrastructure and applications for the cloud. There are a variety of ways to
implement private clouds, which usually involve coupling virtualization with management
and more precisely allocating frameworks that are optimized for pooling virtual resources. The implementation of private
the resources needed to clouds also usually requires tools for enabling self-service provisioning, whereby users select
the services they need, after which the virtualized back-end resources needed to support the
support these services, services are automatically assembled.
private clouds can introduce
significant cost savings and Virtualizing Critical Workloads
improve an organization’s Even before the emergence of cloud computing, virtualization was having a major impact
across the IT industry. In a variety of real-world environments, virtualization has already
responsiveness to changing proven its ability to deliver some key business benefits, including:
business conditions. » Consolidation and improved resource utilization
» Simplified resource provisioning
» Simplified high availability (HA) and disaster recovery (DR)
» Legacy application support
» Improved test and development processes
Consolidation is perhaps the most widely applied use case for virtualization. Consolidation
enables administrators to reduce the number of physical machines that they have to acquire
and manage. The improved resource utilization reduces the server and storage hardware
footprint, which can lower acquisition costs and also reduce some operational costs related to
maintenance, cooling, and power consumption.
Compared to physical servers, virtualization also dramatically reduces the time required to
provision new systems. While the end-to-end time required to bring up a new physical server
can span weeks or months from planning to actual deployment, virtual machines can be
launched in minutes or seconds. Virtualization also makes it easier to reconfigure systems in
response to fluctuating workloads. Live migration of virtual machines between servers provides
flexibility for assigning computing resources to workloads. This capability enables computing
resources to be treated collectively as virtual infrastructure that can be reallocated to different
workloads on demand without having to reconfigure physical resources.
While much of the industry’s recent interest in virtualization has focused on industry-standard
x86 servers, the use of virtualization in fact goes far beyond x86 systems. After being a core
feature of mainframe and UNIX servers for decades, virtualization technology is now available
on all of the major server platforms in use today. Some non-X86 platforms can offer significant
added value for users who want to virtualize enterprise computing workloads (meaning
workloads that may involve multisite, multidepartmental operations, in which disruption can
significantly affect revenue or key business functions). In these environments, users may need
to virtualize very large workloads, or very fast-growing workloads, and they cannot afford to
have hardware or software act as limiting factors to business operations.
Cloud computing further increases the need for robust virtualization platforms, as the
virtualization layer effectively becomes the central point of control for all computing resources.
Therefore, platforms that need to support enterprise cloud computing deployments require a
variety of optimizations to maximize the scalability and reliability of virtualized workloads.
2 WHITE PAPER | 2012 IDEAS INTERNATIONAL, INC.
3. JANUARY 2012 SEEDING CLOUDS ON POWER SYSTEMS WITH IBM SMARTCLOUD™ ENTRY
Scalable Virtualization
One of the main reasons Performance and scalability are two of the greatest priorities for administrators when deploying
enterprise workloads. In an enterprise environment, users need to have confidence that their
for using virtualization to
infrastructure will be able to deliver acceptable service levels even under heavy or rapidly
implement clouds is that changing load conditions. The success of any cloud computing platform that depends on
it can be used to make virtualization will thus hinge on its ability to deliver whatever resources are needed by a
workload, whenever those resources are needed. The scalability of a virtualization platform that
sure workloads always has to support clouds can be measured in two ways:
have sufficient resources » Host scalability. The scalability of the host environment in a virtualization platform
determines how many virtual machines can be hosted on a single server, and thus
to provide the necessary determines the scale of consolidation that can be achieved, and the economies of scale that
service levels, even under can be achieved in a cloud environment. There are three factors that determine the host
scalability of a hypervisor:
variable demand.
– The largest amount of physical memory that is usable by the hypervisor
– The largest symmetric multiprocessing (SMP) configuration of server hardware that the
hypervisor can exploit
– The maximum number of virtual machines that can be deployed on a single server
» Guest scalability. The scalability of a guest environment determines which classes of
workloads are suitable for virtualization on the platform. Guest environment scalability
primarily depends on two limits:
– The largest amount of memory that is accessible by guest operating systems running in
the virtual machine
– The ability for the virtual machine to provide multiple virtual processors that can be used
simultaneously in virtual SMP configurations
Reliable Virtualization
The reliability of infrastructure also remains one of the greatest concerns in enterprise
environments. Cloud computing inherently involves some degree of centralization, which is a
key enabler for achieving the economies of scale that deliver much of the cloud’s value. Using
virtualization for consolidation enables fewer physical servers to be deployed, which reduces
the footprint for potential hardware failures. However, consolidation also increases the
importance of avoiding downtime in individual servers, both unplanned and planned. Therefore,
virtualization platforms that have to support cloud computing need to address reliability at
several levels, including:
» Fault resiliency. For virtualization platforms to handle the most critical workloads, they
require the ability to adapt to certain failures that can occur in electronic components such
as memory, CPUs, and I/O interfaces. Although server hardware has gotten more reliable
over time, restarting large numbers of virtual machines on backup hardware in the wake of
a server failure can still be time consuming and disruptive. Therefore, host failure should be
avoided to the greatest extent possible.
» HA clustering. Virtualization simplifies recovery from failures. An entire virtualized workload
can easily be relaunched simply by restarting the virtual machine on which it is hosted.
Further, when coupled with HA clustering functions, virtualization can be used to
automatically restart workloads on a backup host in the wake of a primary host failure –
dramatically simplifying the implementation of DR procedures.
» Live migration. The ability to migrate virtual machines from one host to another with little or
no interruption to processing is a key benefit of virtualization that helps to reduce planned
downtime. Live migration functions of a virtualization platform should allow workloads to be
temporarily moved so that routine hardware maintenance can be performed on the hosts
with minimal disruption.
WHITE PAPER | 2012 IDEAS INTERNATIONAL, INC. 3
4. SEEDING CLOUDS ON POWER SYSTEMS WITH IBM SMARTCLOUD™ ENTRY JANUARY 2012
The scalability of a virtualization host platform also plays an important role in maintaining
The primary challenge for availability of virtualized workloads. One of the main reasons for using virtualization to
implement clouds is that it can be used to make sure workloads always have sufficient
most users is to stage the resources to provide the necessary service levels, even under variable demand. To maximize
gradual adoption of private consolidation ratios and make optimal use of real resources such as space, power, and cooling,
“scale up” configurations of SMP server hardware should be exploited to achieve as much
clouds in a way that allows performance as possible within a single server footprint. Virtualization platforms should also
the benefits of cloud comput- have the ability to dynamically and quickly reallocate resources to virtualized environments,
including adding and removing CPU, memory, and I/O resources in response to administrator
ing to be accrued as quickly command or workload management software.
as possible. The platform
that customers select as Virtual Infrastructure
they begin the journey to the As the use of virtualization becomes a part of standard operating procedure across the
industry, users are developing strategies to drive virtualization technology more directly into IT
cloud becomes instrumental infrastructure, extending the scope of its functionality from single servers to multiple systems
in meeting this challenge. throughout a datacenter or organization. Using cloud computing for IaaS usually depends on
the deployment of such virtual infrastructure, in which multiple virtual machine hosts can be
treated as a single, flexible pool of resources that are allocated dynamically in response to
changing workload conditions or downtime events. Virtual infrastructure is enabled in part by
the ability to perform online migration of virtualized workloads, whereby the workloads are
migrated from one host to another with minimal disruption.
With virtual infrastructure, the virtualization platform becomes the primary gateway between
workloads and external storage and networks. The virtualization platform thus carries a
fundamental responsibility for moving data between virtualized applications and the real world.
Therefore, the storage and networking capabilities of a virtualization platform become
especially critical components in organizations that are virtualizing applications with heavy I/O
demands. Virtualized storage and networking are also key concerns for organizations that are
virtualizing entire datacenters or computing infrastructures, requiring multiple virtual machine
hosts to share data and processing over the network.
Finally, once virtual infrastructure is in place, users require tools to manage virtual resources
and control how they are assigned to workloads. Because virtual machines inherently decouple
workloads from the details about the hardware on which they are deployed, administrators gain
significant flexibility in matching applications with their required computing resources. However,
the addition of a virtualization layer within systems potentially introduces uncertainty – and
complexity – because the relationship between workloads and computing resources inherently
becomes indirect. Management tools should thus provide visibility over both the physical and
virtual layers of infrastructure, preferably from the same interface. If possible, administrators
should also have consistent visibility and control over other layers of infrastructure, such as
storage and networks. The closer the relationship between the management of all these
components, the easier it will be for administrators to achieve the operational benefits that
cloud computing promises.
Making the Leap from Virtualization to Cloud Computing
As described above, the deployment of private clouds has some basic technical prerequisites,
including robust virtualization platforms with effective virtualization management tools.
However, in order to successfully deploy clouds, IT organizations must also adopt certain
operational practices for managing virtualized systems. These include:
» Simplifying IT services, and standardizing on a select set of virtualized images to implement
these services
» Implementing robust processes for managing the lifecycle of these virtualized images
4 WHITE PAPER | 2012 IDEAS INTERNATIONAL, INC.
5. JANUARY 2012 SEEDING CLOUDS ON POWER SYSTEMS WITH IBM SMARTCLOUD™ ENTRY
» Monitoring the behavior of virtualized workloads and tracking the internal consumption of
The combination of PowerVM resources by user or workload
and IBM Systems Director » Automating provisioning tasks
enables the virtualization Once these practices are in place, it becomes possible to implement the self-service driven
of enterprise workloads that catalogs associated with cloud computing. These catalogs allow users to rapidly deploy their
own services with a minimal knowledge about the underlying computing infrastructure. By
have the highest demands eliminating much of the administrative overhead required to deploy services, private clouds can
for scalability and reliability, introduce significant cost savings.
and it can serve as a solid There are many possible paths for embracing cloud computing in the manner described above.
foundation to adapt such It is essential that organizations choose the approach that matches their specific requirements,
both in terms their ability to absorb cloud techniques, as well as the ability to align service
workloads for cloud levels with workload priorities. The primary challenge for most users is to stage the gradual
computing. adoption of private clouds in a way that allows the benefits of cloud computing to be accrued
as quickly as possible. The platform that customers select as they begin the journey to the
cloud becomes instrumental in meeting this challenge.
Virtualizing Critical Workloads with IBM PowerVM
IBM offers three strategic server platforms, each of which can play a role in the deployment of
cloud computing. One of these platforms, IBM Power Systems, has recently seen a significant
increase in interest as a foundation for hosting clouds. Power Systems include a complete set
of virtualization capabilities that are optimized for the POWER® processor, called PowerVM.
When coupled with the necessary software functions to implement clouds, PowerVM provides
an optimal platform on which to pool computing resources that workloads can draw upon as
needed, while allowing users to activate services on a self-service basis.
There are several reasons why users may be considering PowerVM as a platform on which to
host private clouds. First, Power Systems rank among the most scalable server platforms in the
industry. The POWER7® processor used in Power Systems has outstanding per-core
performance, and servers can scale up to 1,024 virtual cores and 8 TB of memory (on the
largest system, the Power 795). Power Systems currently hold four of the top five nonclustered
(that is, SMP based) rankings on the TPC Benchmark™ C (TPC-C),1 which is one of the most
demanding validations of SMP server architectures in the industry, putting maximum stress on
processors, memory, and I/O. Thus, PowerVM has proven its capability as an extraordinarily
scalable platform on which computing resources can be pooled with confidence.
Security and compliance are vital to many businesses, especially now that they must adhere to
regulatory requirements designed to safeguard personal data and company information from
attacks and breaches. Ensuring that cloud deployments are compliant with common industry
security standards and maintaining system security can be a challenging, labor-intensive
activity especially with today’s virtualized IT infrastructures. IBM® Power Security and
Compliance (PowerSC™) provides a security and compliance solution optimized for virtualized
cloud environments on Power Systems servers, running PowerVM and AIX®.
Further, Power Systems offer very high levels of availability. The high degree of integration in
the POWER7 processor and its advanced multichip packaging enable a number of mainframe-
class reliability functions. Power Systems have a variety of self-management and self-healing
capabilities, such as First Failure Data Capture and Clustering Data Aggregation Tool, Chipkill
with bit-steering/redundant memory, deallocation of PCI bus and L2/L3 cache lines, and
1
As of January 2012, three of the top-five nonclustered TPC-C results are on IBM Power Systems, and
one is a POWER7-based system from Bull, which is tied for first place.
WHITE PAPER | 2012 IDEAS INTERNATIONAL, INC. 5
6. SEEDING CLOUDS ON POWER SYSTEMS WITH IBM SMARTCLOUD™ ENTRY JANUARY 2012
persistent memory deallocation. These functions allow a Power Systems server to continue
IBM SmartCloud Entry is a running even in the event of certain hardware failures. The First Failure Data Capture
mechanism can be used alone on single servers, or configured to work on multiple systems
solution that enables users using a framework called Clustering Data Aggregation Tool. Clustering Data Aggregation Tool
to rapidly deploy private enables more rapid analysis and root-cause identification for problems in workloads that span
multiple operating system instances. IBM complements the reliability and availability of the
clouds based on their exist- Power Systems platform with highly functional and cost-effective DR solutions such as
ing Power Systems infra- PowerHA®, PowerHA SystemMirror® for AIX Standard Edition, and PowerHA SystemMirror for
AIX Enterprise Edition. These options represent some of the most proven and functionality-rich
structure. SmartCloud Entry solutions in the industry for HA clustering and DR.
allows administrators to
The virtualization functions in PowerVM are based on Micro-Partitioning®, which enable many
deploy a self-service portal distinct workloads to share a POWER7 processor and memory simultaneously while running on
on top of PowerVM and IBM different operating systems. (Power Systems support three operating systems: AIX, IBM i, and
Linux, all of which can be virtualized with PowerVM.) The superior scalability of Power Systems
Systems Director VMControl applies directly to the scalability of PowerVM itself. PowerVM can support up to 256 virtual
Enterprise Edition. processor cores (representing 1,024 threads) in a partition, allowing very large workloads to be
virtualized just as if they ran on physical hardware. Operating systems that run in PowerVM,
including AIX and IBM i, have the ability to add and remove resources such as CPU, memory,
and I/O without being rebooted. This capability allows the boundaries between partitions to be
adjusted dynamically in response to changing workloads without interrupting operations.
PowerVM’s Virtual I/O Server (VIOS) allows a single Micro-Partition to provide I/O connectivity
for the other Micro-Partitions in a server and eliminates the need for I/O devices (storage and
network adapters) in each partition, reducing cost and complexity. VIOS partitions can be
configured in redundant pairs, so that each can take over the operations of the other in the
event of failure.
Some of the other important virtualization functions provided by PowerVM include the
following:
» POWER Hypervisor™ can automatically adjust processor and memory resources in a Micro-
Partition based on current utilization and user-defined priorities (the size of a Micro-Partition
can be adjusted in increments of 1/100 of a processor).
» Active Memory™ Sharing improves utilization of system memory by allowing physical
memory in a machine to be assigned into a shared pool, and allocated to partitions on the
fly. Active Memory Sharing also includes support for memory overcommitment, so that if all
partitions in the physical server request more virtual memory than is physically available, the
partitions can use the VIOS partition as a paging device.
» Shared Dedicated Capacity and Multiple Shared Processor Pools can be used to dynamically
and automatically allocate resources to Micro-Partitions based on business-driven policies.
» Live Partition Mobility allows a partition to migrate from one physical host to another without
interrupting the processing of its workloads. This function is essential for reducing planned
downtime, since it allows servers to be taken offline for maintenance while allowing
virtualized workloads to continue providing service. It also increases flexibility for
administrators by allowing them to match virtualized workloads with the resources they
need, regardless of which physical server those resources happen to be available on.
Virtualization Management on IBM Power Systems
All IBM server platforms can be managed through the same interface using a tool called IBM
Systems Director, which is optimized to measure and control low-level functions in IBM’s
servers, including power draw. Systems Director can be used to manage environments with a
mix of physical and virtual systems from a single, consistent interface. To help manage
virtualized environments, Systems Director provides a set of plug-ins, called Systems Director
6 WHITE PAPER | 2012 IDEAS INTERNATIONAL, INC.
7. JANUARY 2012 SEEDING CLOUDS ON POWER SYSTEMS WITH IBM SMARTCLOUD™ ENTRY
VMControl™, for performing specific virtualization management tasks. VMControl is available in
SmartCloud Entry gives users the following editions:
a quick and easy way to set » VMControl Express Edition can be used to create, modify, and delete virtual machines, or
trigger the live migration of virtual machines from one host to another.
up private clouds, while fully » VMControl Standard Edition adds more powerful functions for performing virtual machine
capitalizing on the reliability relocation; importing, editing, creating, and deleting virtual images; maintaining virtual
images in a repository; and deploying virtual images.
and scalability benefits of the
» VMControl Enterprise Edition can be used to create pools of virtualized resources (both
Power Systems platform. server and storage pools). Workloads can tap into these pools on demand in response to
changing conditions.
Another extension for IBM Systems Director, the Storage Management plug-in, can be used to
coordinate virtual machine provisioning and live migration for server, storage, and network
resources. The plug-in can be used to perform lifecycle management of storage systems,
including discovery, health and status monitoring, configuration, updates, and management of
storage virtualization functions.
The combination of PowerVM and IBM Systems Director enables the virtualization of enterprise
workloads that have the highest demands for scalability and reliability, and it can serve as a
solid foundation to adapt such workloads for cloud computing. However, some additional tools
are needed to let users treat a virtualization platform as a cloud. These tools should give users
the ability to rapidly deploy their own services with minimal knowledge about the underlying
computing infrastructure.
Seeding Clouds on IBM Power Systems
As interest in cloud computing started to gain traction across the industry, IBM responded by
assembling a range of cloud deployment solutions based on its systems, software, and services.
These solutions are tailored to meet the requirements of different classes of users, depending
on their progress toward standardization of IT services, their need for integration, and their
interest in retaining control of infrastructure. IBM’s overall strategy for cloud computing is
based on a common architecture called IBM SmartCloud, which targets the following four
classes of cloud usage:
» Business process as a service (BPaaS)
» Software as a service (SaaS)
» Platform as a service (PaaS)
» Infrastructure as a service (IaaS)
To meet customer needs for the latter two classes of cloud deployment, PaaS and IaaS, IBM
has introduced two initiatives:
» IBM SmartCloud Foundation is a set of technologies for deploying private and hybrid clouds
for IaaS. These technologies are designed to let customers easily build and rapidly deploy
private cloud environments based on integrated solutions from IBM, as well as IBM
management tools.
» IBM SmartCloud Services is IBM’s approach to PaaS, whereby IBM provides a service
delivery platform for rapidly deploying applications with maximum scalability, security, and
portability. The initiative includes an array of offerings that allow customers to host a variety
of specific application services on platforms managed by IBM and its partner ecosystem.
WHITE PAPER | 2012 IDEAS INTERNATIONAL, INC. 7
8. SEEDING CLOUDS ON POWER SYSTEMS WITH IBM SMARTCLOUD™ ENTRY JANUARY 2012
FIGURE 1 Users who plan to deploy private clouds will be most interested in IBM SmartCloud Foundation.
With this set of technologies, IBM foresees private clouds being deployed in three possible
IBM SmartCloud Entry Management
stages:
Stack Architecture with PowerVM
» Consolidation and virtualization are the first steps toward implementing private clouds based
• Approval policy
• Project management
on IaaS. At this stage, workloads are rehosted as system images in virtual machines, which
• Users and roles are deployed on servers equipped with robust virtualization capabilities. These virtual
• Events and auditing
• Metering machines can be securely stacked on physical hardware, allowing computing resources to be
• Cloud image library
assigned to workloads with far greater precision, which leads to better utilization of servers.
IBM SmartCloud Entry Management Stack
» Basic cloud infrastructure is the next level of private cloud deployment, which becomes
possible after some standardization of IT services has occurred. At this stage, virtual
Self-Service User Interface
infrastructure can be enhanced with functions that automate certain management
REST API processes, and users can be given some ability to perform self-service provisioning and
Cloud Administration and Management administration.
» Advanced cloud infrastructure is the final stage of private cloud deployment, when
IBM Systems Director VMControl computing resources can be accessed on demand from an integrated and secure virtual
VMControl API
infrastructure. At this stage, the cloud platform has complete support for monitoring the use
of resources, planning capacity, and performing ongoing analytics. Metering and billing
IBM Systems
IBM Systems Director functions can be used for tracking consumption of shared virtual computing resources and,
Director VMControl
where appropriate, allocating costs to users on a chargeback basis. In sufficiently advanced
clouds, the applications themselves may also be optimized for cloud deployment.
PowerVM
IBM recognizes that many users are not ready to skip immediately to the deployment of
OS OS OS advanced private clouds, despite their potential benefits. More likely, depending on their culture
and priorities, organizations will begin to adopt cloud computing with simpler initiatives, and
IBM Power Systems progress gradually to more sophisticated cloud deployments. Correspondingly, IBM has started
to refine its cloud offerings in order to give customers a choice of entry points.
Customer-specific integration
At the high end, IBM SmartCloud Foundation has offerings that range from high-touch service
engagements, in which IBM will take on the entire task of bringing cloud computing to a
customer based on their specific needs, to highly integrated solutions that customers can install
themselves, such as IBM Service Delivery Manager. These solutions generally provide complete
capabilities for service orchestration and integration, automated IT service deployment, end-to-
end lifecycle management, and metering and chargeback.2
However, the high-end solutions in IBM SmartCloud Foundation can incur considerable up-front
investments to plan and deploy, and they may also require users to take on some complexity in
order to achieve the complete benefits of clouds. Some customers may seek a simpler
approach that allows them to deploy clouds quickly, and without major changes to operations.
For these users, IBM has introduced an offering called IBM SmartCloud Entry, a scaled-down
software package that enables clouds to be deployed rapidly and with minimal disruption to
existing infrastructure and administrative processes.
IBM SmartCloud Entry is a solution that enables users to rapidly deploy private clouds based on
their existing Power Systems infrastructure. SmartCloud Entry allows administrators to deploy a
self-service portal on top of PowerVM and IBM Systems Director VMControl Enterprise Edition
(see Figure 1, upper left). With a flexible, modular design and a small footprint, SmartCloud
Entry offers the following features:
» An administrative interface to create and replicate cloud services, as well as manage the
deployment of these services and monitor their activity.
2
See Private Clouds Float with IBM Systems and Software, Ideas International, February 2011 at:
http://public.dhe.ibm.com/common/ssi/ecm/en/xbl03006usen/XBL03006USEN.PDF.
8 WHITE PAPER | 2012 IDEAS INTERNATIONAL, INC.
9. JANUARY 2012 SEEDING CLOUDS ON POWER SYSTEMS WITH IBM SMARTCLOUD™ ENTRY
» A library for storing standardized images containing virtualized AIX and Linux workloads. The
With SmartCloud Entry, users library allows administrators to easily create new “golden master” images and software
appliances. These images can be based on an organization’s standard configurations of
can start to set up the admin- operating systems, applications, technology levels, and security patches. The management
istrative processes that are interface also allows administrators to track the configuration and usage of images to ensure
compliance and minimize security risks.
necessary for the complete
» A self-service request catalog that allows provisioning to be delegated to authorized users.
transition to cloud computing, By enabling certain users to provision virtualized workloads themselves, with little or no
which include standardizing administrator intervention, the catalog helps to improve management productivity.
» Project-level user access controls to limit resources that can be assigned to particular
IT services, developing pro- workloads, enforced by the underlying virtualization capabilities of PowerVM.
cedures for users to perform » Basic workload metering to support a “pay-per-use” business model, with a single interface
self-service provisioning, and for analyzing and managing approvals, metering, billing, users, and projects.
» REST APIs that allow the management platform to be extended with custom-designed user
in some cases, introducing interfaces.
resource metering and billing.
The integration of SmartCloud Entry with IBM Systems Director VMControl Enterprise Edition
allows servers and storage to be assigned to pools of virtualized resources that virtualized
workloads can tap into on demand. As a result, cloud administrators can quickly scale up
virtualized infrastructure by simply adding servers, blades, or storage devices to these pools.
SmartCloud Entry gives users a quick and easy way to set up private clouds, while fully
capitalizing on the reliability and scalability benefits of the Power Systems platform. As these
users become more comfortable with cloud computing, and seek to gain additional cost and
responsiveness benefits from increased integration, they can progress from SmartCloud Entry
to IBM’s more advanced cloud offerings, such as IBM Service Delivery Manager. In the
meantime, SmartCloud Entry offers a relatively low-risk method of introducing the key
operational processes required for cloud computing (standardizing IT services, building libraries
of these services, and implementing procedures for users to provision these services on their
own).
The IDEAS Bottom Line
Many organizations are genuinely excited about the potential for private clouds, anticipating
greater responsiveness for end users and reduced costs for administrators. These organizations
may have already achieved considerable success in their adoption of virtualization, and they
would now like to build on that success with extensions for enabling private clouds. However,
they are also concerned with minimizing the risk of disruption as they make the transition to
cloud computing, which by definition introduces new management processes that
fundamentally redefine how resources are assigned to workloads.
One way to reduce the risk of deploying private clouds is to make the simplest possible
extensions to existing virtualization platforms for adding key cloud capabilities. These
extensions should include tools for creating service catalogs, and self-service portals that allow
users to assume some self-service provisioning. The tools can offload some management tasks
from administrators, and thus help to reduce management costs. Users may achieve a
relatively quick return on investment from private clouds by deploying them on server
platforms with characteristics that already align well with the core requirements of cloud
computing. These characteristics include sufficient scalability to support massive levels of
consolidation, and the resiliency to maintain uptime for services in the event of hardware or
software failures.
IBM SmartCloud Entry provides a set of building blocks that will allow users to deploy private
clouds quickly and easily on the proven virtualization capabilities of IBM Power Systems. With
WHITE PAPER | 2012 IDEAS INTERNATIONAL, INC. 9
10. SEEDING CLOUDS ON POWER SYSTEMS WITH IBM SMARTCLOUD™ ENTRY JANUARY 2012
About Ideas International (IDEAS) SmartCloud Entry, users can start to set up the administrative processes that are necessary for
IDEAS provides enterprise IT the complete transition to cloud computing, which include standardizing IT services, developing
research, insight, analysis, and tools procedures for users to perform self-service provisioning, and in some cases, introducing
to computer suppliers and consultants resource metering and billing. Once these processes have matured, organizations can proceed
(IT Sellers) and large corporations (IT to more advanced cloud infrastructures with stronger integration between servers,
Buyers). The company’s research management software, and ultimately, applications. By deploying SmartCloud Entry on IBM
focus areas include servers, storage, Power Systems, organizations can take the first steps toward this goal with a relatively modest
software, services, and cloud. Many investment. Once these initial steps have been completed, the superior scalability and
IDEAS tools are powered by RPE2, availability of Power Systems will enable users to achieve ever-increasing economies of scale in
their IT infrastructure, without having to switch platforms as they apply cloud computing
the atomic unit of compute. IDEAS is
practices more and more broadly.
a publicly traded company on the
Australian Stock Exchange (ASX:IDE)
and has been in business for over 25
years. IDEAS hosts users in over 100
countries and maintains offices in the
US, EMEA, and Asia Pacific.
Americas
Ideas International, Inc.
800 Westchester Avenue
Suite N337
Rye Brook, NY 10573-1354
USA
Tel + 1 914 937 4302
Fax +1 914 937 2485
Asia/Pacific and Worldwide
Headquarters
Ideas International Limited
Level 3
20 George Street
Hornsby, NSW, 2077
Australia
Tel +61 2 9472 7777
Fax +61 2 9472 7788
Europe, Middle East, Africa
Ideas International Europe
Milton Park Innovation Centre
99 Milton Park
Abingdon, Oxon OX14 4RY
United Kingdom
Tel + 44 (0) 1235 841 510
Fax + 44 (0) 1235 841 511
actionable intelligence
www.ideasinternational.com
10 WHITE PAPER | 2012 IDEAS INTERNATIONAL, INC.