-
Understanding the HP CloudSystem Matrix Technology
Technical white paper
Table of contents
Introduction..................................................................................................................................2
Overview: HP CloudSystem Matrix
.................................................................................................3
How CloudSystem Matrix is used
...................................................................................................4
Designing and provisioning infrastructure services
........................................................................5
Optimizing the infrastructure for capacity planning and
showback/chargeback .............................13 Protecting
service continuity with automated cost-effective failover and
disaster recovery .................16
Enabling technology
...................................................................................................................17
BladeSystem c-Class
................................................................................................................17
Virtual
Connect.......................................................................................................................18
Matrix Operating
Environment..................................................................................................20
Multi-tenancy features in Matrix
OE...........................................................................................20
Cloud Service Automation for
Matrix.........................................................................................21
Storage technologies
...............................................................................................................21
Security
.................................................................................................................................22
Integrating HP CloudSystem Matrix into customer environments
....................................................22 Expanding
beyond IaaS: CloudSystem portfolio
.........................................................................24
Purchase and delivery
.................................................................................................................25
HP services to make the most of
CloudSystem.............................................................................27
Ready capacity when business demands with bursting and
pay-as-you-use resources ......................28
Summary and
conclusion.............................................................................................................29
Appendix A: Implementing an IaaS
..............................................................................................30
Appendix B: HP CloudSystem Matrix use
cases..............................................................................30
Move your Test & Development environment to a Cloud
..............................................................30
Expand the Cloud to include your production infrastructure
..........................................................30
Deliver and monitor both infrastructure and
applications..............................................................30
Appendix C: HP Cloud Mapsan
example...................................................................................31
Glossary of
terms........................................................................................................................34
For more
information...................................................................................................................37
-
2
Introduction Businesses are moving to the cloud in accelerating
numbers. Business users recognize cloud advantages that help speed
innovation, accelerate business processes, and reduce time to
revenue. Frequently, these users turn to public cloud services for
the solutions they need. In fact, a Forrester study found that
business is adopting cloud 2.5 times faster than is IT operations.1
IT directors, faced with the challenges associated with the rush to
cloud-based services, are concerned: how can IT maintain adequate
security, ensure service levels, integrate cloud-based services
with existing traditional IT systems, and provide governance across
the entire IT value chain? In addition, IT directors want to be
sure that the decisions they make today about cloud technology
suppliers dont prevent them from innovating in the future.
At the same time, IT departments are under tremendous pressure
to keep pace with business demand. While IT infrastructure has made
significant advances in performance and availability, the effort to
meet increasing business demands has frequently led to IT sprawl.
Computing resources proliferate, but remain underutilized and too
hardwired to redeploy easily when business needs change. IT
increasingly employs virtualization and automation to improve the
flexibility and increase utilization of computing resources.
However, this course has not yet eliminated the issues of
over-provisioning and complexity. Too frequently, IT resources are
dedicated to a particular application or business unit and any
excess capacity remains unavailable for other uses. IT continues to
be hampered in its ability to focus its resources on strategic
objectives and driving innovation, resulting in slow time to
revenue and lost opportunity.
IT departments are evolving from the role of sole supplier to
becoming both builder and broker of IT services. As such, IT must
build its capability as an internal service provider that matches
the transparency and flexibility of externally available services,
as well as be able to source and consume services from a network of
trusted suppliers.
HP offers you a path to providing secure, predictable
cloud-based services through Converged Infrastructure. Your IT
staff can consolidate physical and virtual server, storage, and
network assets into pools of virtualized resources. These resource
pools can host sets of infrastructure services, which typically map
to application services, including complex multitier, multi-node
applications. IT personnel can flexibly provision and re-provision
services, and can confidently optimize the underlying resources for
performance, resiliency, and efficiency.
HP Converged Infrastructure technologies are at the core of the
HP CloudSystem portfolio: a comprehensive, integrated, and open
solution that provides IT with a unified way to provision and
manage services across private, managed, and public clouds. HP
CloudSystem equips you to respond to customers or business units
faster, more predictably, more efficiently, and with lowered costs.
HP CloudSystem offerings provide a range of services as well as an
avenue for growth and expansion. As shown in Figure 1, HP
CloudSystem has three integrated offerings.
HP CloudSystem Matrix: Enables you to deliver infrastructure as
a service (IaaS) and provisions infrastructure and basic
applications in minutes.
HP CloudSystem Enterprise: For those looking to deploy private
and hybrid cloud environments and the full range of service models
(IaaS, platform as a service (PaaS), and software as a service
(SaaS). This offering provides a single service view of your
environments, from private cloud to public clouds to traditional
IT, with advanced application-to-infrastructure lifecycle
management.
HP CloudSystem Service Provider: Public or hosted private cloud
designed for service providers to provide a public cloud IaaS and
SaaS, including aggregation and management of those services.
1 Forrester Research, Inc., Ignoring Cloud Risks A Growing Gap
Between I&O And The Business, March 24, 2011
-
Figure 1: Three integrated HP CloudSystem offerings for a full
range of cloud capabilities
This whitepaper describes the HP CloudSystem Matrix solution,
with a focus on how it delivers IaaS. The document is intended for
IT directors, IT architects, solution architects, and other readers
who are familiar with current HP BladeSystem offerings and existing
server virtualization technology. The content should assist you in
understanding how HP CloudSystem Matrix can benefit your IT
environments with more flexible, efficient use of physical and
virtual IT resources. The paper begins with a discussion of key
architectural and operational concepts and then highlights the key
enabling technologies; it concludes with a brief overview of the
purchase and implementation process, including available
services.
Overview: HP CloudSystem Matrix HP CloudSystem Matrix is an IaaS
solution for private and hybrid cloud deployments, built on proven
HP Converged Infrastructure technologies, such as HP BladeSystem,
Matrix Operating Environment and Cloud Service Automation for
Matrix. HP CloudSystem Matrix allows you to:
Provision infrastructure and applications in minutes for
physical and virtual environments. Reduce TCO up to 56%2 with
built-in infrastructure life-cycle management. Integrate
heterogeneous environments into your IaaS infrastructure. HP Cloud
Maps accelerate automation of cloud service deployments and ensure
consistency and reliability of the implementation of infrastructure
service catalogs.
CloudSystem Matrix is an integrated hardware, software, and
services solution that helps you realize the full value of cloud
computing as quickly as possible. The core elements of a
CloudSystem Matrix solution are:
HP BladeSystem c7000 enclosures (one or more) HP Virtual Connect
HP Matrix Operating Environment HP Cloud Service Automation (CSA)
for Matrix HP Implementation Service
2 Compared to traditional infrastructure, based on analysis with
the BladeSystem and BladeSystem TCO calculators:
http://www.hp.com/go/matrixtco
3
-
HP CloudSystem Matrix is optimized for HP ProLiant and HP
Integrity servers as well as HP storage and HP networking, but also
supports 3rd party x86 servers, networking, and storage. Supported
operating systems are Windows, Linux and HP-UX. Supported
hypervisors include VMware, Microsoft Hyper-V, and Integrity
VMs.
The Matrix Operating Environment includes an integrated service
designer, self-service infrastructure portal, and auto-provisioning
capabilities. It includes the tools to manage and optimize resource
pools, multi-tenancy, and a recovery management solution for
ProLiant. Recovery management for HP-UX is available with HP
Serviceguard. Using included Cloud APIs, you can easily customize
the operating environment to your specific requirements, enabling
chargeback and billing integration, integration into approval
processes, and other process automation tasks. CSA for Matrix
integrates basic application provisioning, monitoring, and patch
and change management. And because it is vital to get started down
the right path, CloudSystem Matrix also includes a comprehensive
Implementation Service. This service is performed by HP trained
experts, including a project manager to manage the technical
preplanning, installation, configuration, testing, demonstration,
and orientation of your integrated Matrix package solutions.
There is also an upgrade path for future HP BladeSystem
purchases as IT needs grow, all while protecting the existing HP
BladeSystem investment. The HP CloudSystem Matrix Conversion
Services provide a cost-effective way to start realizing the
benefits of shared services and a private cloud today. CloudSystem
Matrix is available as a small, medium, or large configuration,
which can be expanded with additional hardware and software from HP
or third parties.
Figure 2 provides a diagram of the HP CloudSystem Matrix
solution with Virtual Connect FlexFabric.
Figure 2: Solution diagram of HP CloudSystem Matrix with Virtual
Connect FlexFabric module*
* Optionally, Virtual Connect Flex-10 modules can be used for
LAN connection and Virtual Connect Fibre Channel Modules for SAN
connection.
** Cloud Service Automation for Matrix is an optional component
of the CloudSystem Matrix solution.
How CloudSystem Matrix is used Perhaps the best way to
understand the unique value of the solution is to examine how you
can build a private and hybrid cloud, and deliver and operate
infrastructure services, using HP CloudSystem Matrix and the
capabilities of the Matrix Operating Environment and CSA for Matrix
(illustrated conceptually in Figure 3).
4
-
Figure 3: CloudSystem Matrix automates the entire lifecycle to
build and manage a private and hybrid cloud
The Matrix Operating Environment enables provisioning and
re-provisioning of the shared pools of servers, storage, power, and
network connectivity as needed, based on pre-defined templates.
Typically, each template represents the complete infrastructure
needed to host a specific application service, such as Microsoft
Exchange, Oracle Real Application Clusters (RAC), an enterprise
resource planning solution, or even a custom application, and may
include the application itself. An IT architect uses a graphical
designer tool to build and publish these infrastructure service
templates. Using a self-service infrastructure portal, IT users can
view a catalog of published service templates and request an
instance of a service to be automatically provisioned. Through its
infrastructure orchestration and core multi-tenancy capabilities,
Matrix provides standardization, resource sharing across client
organizations, and efficient provisioning control for a
private/hybrid cloud infrastructure.
Compared to a manual process that requires coordination among
different teams across a data center or IT organization,
CloudSystem Matrix enables administrators to provision an entire
infrastructure service, including the application, in just minutes
or hours instead of weeks or months. Administrators can also
continuously monitor and optimize the cloud infrastructurefor
example, by managing capacity, moving workloads or adjusting
infrastructure service lease periodsas well as protect service
continuity with disaster recovery and automated server failover.
Advanced capabilities enable an IT administrator to provision
services outside of the private cloud managed by Matrix by bursting
into a public cloud to provision template-defined services as
needed. The following sections describe in further detail how
CloudSystem Matrix enables provisioning, optimizing, and protecting
continuity of cloud infrastructure services. (For definitions of
many of the terms used here, see the Glossary of terms.)
Designing and provisioning infrastructure services The most
powerful attribute of HP CloudSystem Matrix is the speed and
simplicity with which pools of infrastructure resources can be
carved up and flexibly configured and reconfigured to match the
infrastructure requirements of nearly any application. You can
deploy a single CloudSystem Matrix environment in place of multiple
different server configurations to satisfy varied application
requirements for CPU, memory, network, and storage and quickly
build your private cloud. Matrix resources are then available on
demand as infrastructure services for IT users to request using a
service catalog.
In a traditional IT environment, the deployment of a new
application requires the involvement of many people and a high
degree of coordination among them. An IT architect might draw up a
design specifying the servers, virtual machines, storage, and
networks needed to support the application. The design would
specify the connections between those resources and adhere to
established policies and standards. Based on this design, the
different IT staff responsible for servers, storage,
virtualization, networking, and facilities would assemble and/or
activate the needed
5
-
resources. This process could take multiple weeks, if not
months, and involve repeated communication between the different IT
groups.
Making use of role-based infrastructure orchestration software,
Matrix enables:
Administrators to create pools of resources and oversee and
control their use. IT architects to design the infrastructure to
support business applications and publish service templates to
drive
implementation.
Users of IT services to select from a service catalog and
request infrastructure services.
The CloudSystem Matrix software simplifies, streamlines, and
coordinates these processes and automates the provisioning of the
infrastructure resources to satisfy approved requests.
Creating and managing resource pools An administrator can use
the management console to define pools of server, network, and
storage resources to support infrastructure services (Figure 4).
Through the console, the administrator can also define virtual
machine images and software deployment jobs, register user
accounts, and assign resource pools to users.
Figure 4: Creating server pools
Storage pool entries describe the storage requirements for the
infrastructure services. The Matrix administrator (typically a
server administrator) can create a storage pool entry to define the
storage needs of a physical or virtual server (for example, a boot
disk of a specific size, a shared data disk, and a private
transaction log disk).
A storage administrator can create a storage catalog
representing the available storage assets. These assets can include
existing storage volumes that have been pre-provisioned by the
storage administrator as well as specifications for volumes that
can be created on demand. Storage templates encapsulate policies
(specifying for example, RAID level, capacity constraints, use of
thin provisioning, and on-demand provisioning if applicable)
associated with the storage volumes. The storage administrator
controls the visibility of the storage volumes and templates in the
catalog and the operations that can be performed.
The storage catalog serves as a formal, automated communication
mechanism in Matrix between Matrix administrators and storage
administrators. It eliminates the need for manually communicating
storage needs and then manually entering the storage information.
Instead, the Matrix administrator can directly generate a storage
request by creating the storage pool entry and then select an
appropriate candidate from the storage catalog. If there are no
6
-
7
matches in the catalog, the storage administrator can see the
unfulfilled request and provision appropriate storage. The storage
administrator can then fulfill the request with the newly
provisioned storage, and the Matrix storage definition
automatically updates to reflect that storage. Using the Matrix
storage catalog as a communication device significantly improves
the efficiency of the interactions between server and storage
staff.
Matrix can also auto-generate storage pool entries based on the
logical disk specifications in the service templates. In this case,
Matrix automatically matches the best candidate from the storage
catalog to the request. (More information on the software that
enables these capabilities is contained in Storage
technologies.)
Designing the infrastructure and creating service templates The
architect uses a graphical designer (shown in Figure 5) to plan and
design infrastructures to fit the needs of business applications.
For example, a design can be a three-tier architecture, with the
database tier running in physical mode and the application and web
tiers configured to run in virtual mode, or it can be a simple
virtual machine. By dragging and dropping and connecting icons
representing the required resources, the architect creates a
template, saved as an XML file, for an infrastructure service. The
architect can also import and export template files. The architect
specifies attributes for the logical resources, such as minimum
memory required, IP address allocation, network/VLAN
characteristics, requirements for physical or virtual servers, and
the software required on the boot disk.
When considering the storage needed for a service, the Matrix
architect has the option to browse storage templates created by the
storage architect, then select and edit an appropriate storage
template for the type of service. If no appropriate choice is
available, or based on architect preference, the Matrix architect
can specify the logical disk attributes manually (including
information such as size, RAID level, and storage tier or other
optional tags). The architect saves each service template to a
catalog for access only by specified, approved users. The
infrastructure orchestration software validates the design and
presents information about any issues so that they can be resolved
before the service template is published.
-
Figure 5: Creating a template using the graphical designer
The CloudSystem Matrix designer interface enables the architect
to:
Design infrastructure service templates for a range of
application needsfrom a simple virtual machine cluster for a web
front end, to a configuration for a test and development
organization, to complex, multi-tier designs for e-shopping,
enterprise resource planning (ERP), or customer relationship
management (CRM) applications.
Take advantage of Matrix core multi-tenancy capabilities to
create and designate separate infrastructure service templates for
individual business groups.
Incorporate established IT policies and standards with assurance
that they will be followed in the provisioning process.
Specify costs to support budget tracking and reporting. Attach
custom workflows that automate key pre- and post-provisioning
tasks. Use workflows and processes to enable smooth transitions
between roles or teams (Figure 6). Example workflows:
Open Request For Change (RFC) tickets for requests in ticketing/
change management systems Pass customer, cost, and lease duration
into a chargeback system Apply patching and compliance policies in
configuration management tools Maintain status in change management
systems when new services are created or changed
8
-
Figure 6: Adding workflows
The architect can design service templates from scratch, modify
previously created templates, or leverage best practice templates.
HP Cloud Maps, developed with key independent software vendors
(ISVs), include standard templates and other resources to aid in
designing and customizing service templates for applications such
as Microsoft Exchange, Oracle RAC, and SAP NetWeaver, and others
(visit www.hp.com/go/cloudmaps).
Included HP Cloud APIs enable the architect to integrate Matrix
with business processes or other IT operations. For example, the
architect can link Matrix provisioning to a company chargeback
system. In designing a template, the architect can associate a cost
with each of the elements of a service; the service template then
displays the overall cost of the service. Using the interfaces
provided, the architect (or administrator) can easily build the
links to enable the chargeback system to capture usage and cost
summaries from CloudSystem Matrix at the desired intervals. (See
also Optimizing the infrastructure for capacity planning and
showback/chargeback.) For more information on Cloud APIs, see the
white paper, Customize your Cloud with HP Matrix Operating
Environment at
http://h20195.www2.hp.com/V2/GetDocument.aspx?docname=4AA3-6357ENW&cc=us&lc=en.
Enabling multi-tenancy CloudSystem Matrix provides multi-tenancy
capabilities to help enterprises share IT resources securely among
separate business groups. Matrix allows a Matrix administrator,
referred to in this context as the Service Provider administrator,
to support multiple clients or tenants by creating a virtualized
instance of Matrix infrastructure orchestration software for each
tenant group, here referred to as an Organization. Each
Organization supports business requirements for separate
infrastructures and services that include computing, storage, and
network resources. Matrix infrastructure orchestration uses these
separate infrastructures to isolate user information securely for
each Organization. The Matrix Service Provider administrator can
provide each Organization access to specific infrastructure service
templates and to certain pools of servers, storage, and
networksincluding single or multiple VLANs (Figure 7).
Administrators at the Organization level can manage and assign only
the resources within their Organization, and Organization users can
request infrastructure services based only on the catalog of
service templates assigned to their Organization.
9
-
Figure 7: Managing Matrix Organizations for a multi-tenant
environment
Extending provisioning to application Through integration with
optional HP CSA for Matrix software, CloudSystem Matrix offers
automated application provisioning, monitoring, and compliance
management. CSA for Matrix, the integration of HP Server Automation
(SA) and HP SiteScope, enables the Matrix architect to select both
the OS and application software needed for a service (Figure 8),
create a single service template that comprehends both
infrastructure and application, and thus include applications in
the service catalog. Then, when a user requests the service, Matrix
provisions the infrastructure including OS (virtual or physical or
both) and application software as a single process.
10
-
Figure 8: Selecting software using SA of CSA for Matrix as a
provisioning engine
Requesting infrastructure and application services An approved
user can initiate the creation of a new service from the published
templates, which may include infrastructure services or
infrastructure and application services. Through the Matrix
self-service infrastructure portal, the user can:
Select a service template from the catalog (example in Figure
9). Select one or more assigned resource pools. View associated
costs for the service and specify a lease period.
Since provisioning infrastructure resources is generally the
domain of IT, the Matrix portal is designed for IT staff as the
typical users. However, line of business users can also use the
portal to request infrastructure services. Users are able to browse
only the service templates that they are authorized to use based on
their functions. For example, members of a test and development
team might see a variety of infrastructure services, while members
of the finance department would see only services needed to support
finance, and human resources (HR) personnel would see only services
needed for HR functions. The self-service portal displays all of
the users service requests, with their status, and provides email
notification when a status changes.
11
-
Figure 9: Template catalog in self-service portal
Approval and provisioning The Matrix software sends an email
notification of pending service requests to the administrator.
Through the management console, the administrator can review each
request, and approve or deny the request. Approval of a request
initiates the provisioning process.
To the infrastructure orchestration software, the components of
each service template are logical objects. A key capability of the
software is allocationthe process of finding the resources that
match the logical objects. After a user submits a service request
through the self-service portal, the Matrix software validates the
needed infrastructure for the service request and performs an
allocation. If the software can locate the resources that match the
template, the allocation is successful; the software reserves the
selected resources and provisioning proceeds.
CloudSystem Matrix performs automated provisioning for both
virtual and physical servers by including physical server and
virtual machine configurations, allocating appropriate SAN volumes,
deploying and customizing the OSand, with CSA for Matrix,
installing the applications. CloudSystem Matrix supports a variety
of mechanisms for deploying OS software. The software source can be
virtual machine templates or HP server deployment software. When
provisioning several virtual machines from the same virtual machine
template, Matrix can utilize linked clones, which speeds up the
provisioning process and reduces the amount of storage
required.
When the storage template associated with an infrastructure
service request calls for on-demand storage provisioning, Matrix
automatically provisions the storageproviding just in time use of
storage resourcesat the point of service creation. Matrix uses the
specifications and policies in the storage template: creating and
presenting the volume and performing SAN zoning in Brocade SAN
environments. When the associated storage template calls for
pre-provisioned storage, Matrix automatically matches the service
request to the appropriate pre-provisioned storage at the time the
compute and network resources are provisioned.
12
-
13
During provisioning, the Matrix infrastructure orchestration
software invokes any associated workflows at their respective
execution points. Because the service template automates and
controls the provisioning, administrators can enforce established
standards and avoid variances due to human error.
Managing provisioned services Matrix notifies both the
administrator and the user who requested the service when
provisioning is complete. The administrator can view status,
progress, and details of all completed and in-progress requests and
take action to resolve failed requests. The administrator can also
modify service infrastructures as required, for example, migrating
workloads between server blades to support proactive maintenance
activities in the physical environment.
The provisioned service is available to the user for the
duration of the specified lease period. A user can manage
provisioned services through the infrastructure self-service
portalsetting the servers to standby to reduce power consumption if
a service is not needed temporarily, resuming the service and
reactivating the servers when needed again. The user can also
easily request a modified service lease period or request
additional servers or storage. When a service is no longer needed,
the user can delete it to make resources available for other
purposes. Administrators can choose one of three options for a
service at the end of its lease period: ignore the service lease
end and notify the user, suspend the service, or delete the
service.
Ongoing monitoring and compliance management using CSA for
Matrix Upon provisioning of a new service, Matrix communicates the
IP addresses of the associated physical and virtual servers to CSA
for Matrix. CSA for Matrix, through the SiteScope component, then
automatically deploys infrastructure monitors for CPU, memory, and
disk utilization of those servers. Easily configurable application,
database, and OS monitors are also available for deploymentwithout
installing any agentsto help ensure the availability and
performance of application as well as infrastructure
components.
The SA component of CSA for Matrix not only enables automated
application provisioning, it also enables ongoing compliance
management of physical and virtual servers and applications. IT
staff can establish a baseline of servers, local storage, and
software and define compliance policies that govern patch levels
and other configuration variables. Administrators can then run
compliance checks and quickly correct configurations that do not
comply with defined standards. CSA for Matrix greatly improves the
speed and consistency of deploying patches and other software
across hundreds of servers simultaneously using best practices and
proven configurations.
Tapping additional resources through cloud bursting
Out-of-the-box dual bursting capability offered with HP CloudSystem
allows enterprises to access additional resources on an as-needed
and pay-per-use basis. The need may be for more processing capacity
to deal with a sudden spike in demand, for a localized datacenter
presence to extend reach or improve performance in a distant
geography, or for specialized capabilities to address compliance or
other unusual service requirements. Dual bursting means that the
additional resources accessed can include both locally hosted
Utility Ready Computing resources offered by HP and public cloud
resources offered by HP and other service providers. CloudSystem
easily handles bursting to locally hosted resources, to public
cloud services, or to any combination of the two.
The Matrix Operating Environment represents resources available
from public clouds as server pools. Users see and work with both
internal and public cloud server pools in the same way. Service
templates can be designed with private or public cloud deployment
in mind. Matrix employs a simple allocation policy when
provisioning resources, selecting the first of the user-selected
pools that can accommodate the service needs. If resources have
been allocated to a public cloud server pool, Matrix sends
provisioning calls to the public provider to deploy an OS image, to
add data disks, and to put the servers onto the appropriate public
cloud network. CloudSystem monitors progress and reports status of
public cloud deployments just as it does for on-premise
provisioning.
You can find more information on HP programs supporting cloud
bursting in the Ready capacity when business demands section of
this document.
Optimizing the infrastructure for capacity planning and
showback/chargeback Integrated capacity planning capabilities keep
infrastructure services provisioned in HP CloudSystem Matrix
running optimally. CloudSystem Matrix enables administrators to
quickly rebalance and repurpose physical and virtual servers to
address changing business priorities. As energy costs continue to
rise, power consumption is an increasingly important consideration
when planning for server consolidations or growth. Energy-aware
capacity planning tools and what-if scenarios help identify optimal
workload placement based on projected resource utilization,
power
-
14
consumption, and other pertinent criteria. On an ongoing basis,
built-in technologies throttle resources when not needed, keeping
power consumption in line with actual server utilization. Dynamic
power saver features can place power supplies in standby to keep
power supply efficiencies above 90% for all normal operating
conditions.
In addition to enabling effective IT planning, Matrix resource
utilization tracking can support business purposes as well. As
noted in Designing the infrastructure and creating service
templates, CloudSystem Matrix includes features to facilitate
integration with external chargeback and/or billing solutions
(although CloudSystem Matrix itself does not include a chargeback
and/or billing mechanism).
APIs, CLIs and customizable workflows provide a means for
external programs to capture and store chargeback/billing
information
Chargeback/billing related customer selectable fields are
provided in the service catalog and user self-service portal. These
can be used to set and track cost and billing codes of services
This detailed white paper, Collecting Usage Data for Showback,
Chargeback, and Billing Purposes at
http://h20195.www2.hp.com/V2/GetPDF.aspx/4AA3-8068ENW.pdf,
describes some tools and techniques that can be used to extract
usage data in a Matrix environment for showback, chargeback, and/or
billing purposes. It describes the services model in Matrix, the
attributes that would be of interest in a cloud environment, and
the APIs and CLIs to extract the related usage data. It also
describes two different architectural approaches for gathering this
data into an external database for report generation.
Ongoing capacity management Using the integrated real-time
capacity planning and workload balancing capabilities of
CloudSystem Matrix, administrators can easily accommodate the
varying resource requirements of an application workload. For
example, varying the resources based on time of the month can
address dramatic increases in peak use during end-of-month
processing or a sales promotion.
The capacity planning capability of CloudSystem Matrix measures
the traditional resource metrics of CPU, memory, networking, and
disk I/O. Matrix also measures power consumption, based on data
collected every five minutes for the Matrix server blades. In the
management console, administrators can view both current
utilization and historical data by the type of resource. When
resource utilization is low for a set of physical or virtual
servers, administrators can evaluate consolidation alternatives
using consolidation planning scenarios.
It is helpful to understand the CloudSystem Matrix approach to
analyzing capacity requirements, which is more sophisticated than
simply determining the maximum memory or CPU utilization. A common
practice in capacity planning is to take simply the peak of the
various loads and use that to determine the maximum required peak
capacity; this method is the sum of peaks. Although it will provide
a robust solution, this method does not take into account the
timing of the peaks of the loads and may end up planning for much
more capacity than is actually used. A more efficient planning
solution, referred to as the peak of sums and easily accomplished
with Matrix, takes into account the timing of the maximum
utilization peaks in the individual loads. By adding together
utilization at each measured interval and then taking the maximum
of the resulting time sequence, the peak of sums method used by
CloudSystem Matrix results in a more accurate measure of the
required maximum resource, thus reducing costs.
Consolidation planning using what-if scenarios When planning for
consolidation, an IT administrator can simulate the placement of
server workloads prior to implementation and compare resources used
under multiple scenarios, taking into account future trends. By
analyzing these planning scenarios, the administrator can determine
optimal workload placement.
With the Matrix scenario editor, the administrator can create a
baseline scenario based on an existing solution and then generate
the desired number of alternative scenarios for comparison. For
each of the alternatives, the administrator can apply what-if
actions, for example, automated system consolidation to virtual
machines. The administrator can also specify configuration
parameters and include growth projections for CPU, memory, disk
I/O, and networking I/O, as well as utilization limits for each of
these resources. The capacity planner compares the alternatives and
provides a recommended solution, including a 5-star headroom rating
and projected utilization. The administrator can then run a
comparison report to determine possible outcomes. For example,
administrators can compare relative improvement in power
consumption from various alternative scenarios, as illustrated in
Figure 10. Details of the consolidation planning scenarios give the
administrator additional relevant data on which to base a decision
regarding the best consolidation solution. The administrator can
then move the workloads to the selected host and free up the server
blade for other applications or power it down to save energy.
-
Figure 10: Consolidation planning scenario comparison report
Additionally, administrators now have the capability to easily
spot the peaks of individual servers and monitor the average
utilization across all servers through a peak summary report. The
administrator can then view a consolidation candidate report and
determine the best candidates for consolidation based on real,
historical data. Finally, a cost allocation report presents
utilization in percentage format, ready for further cost allocation
reporting if desired.
Advanced power management: dynamic power capping The dynamic
power capping feature of Matrix safely limits peak power
consumption without impacting system performance and without risk
of over-subscribing data center branch circuits. This capability is
enabled for every ProLiant server blade and blade enclosure.
Administrators can set power caps for individual server blades, for
groups of blades or for an entire enclosure. Embedded management
and power microprocessors on each blade work together both to
measure and control power usage. When enforcing the user-defined
power cap, the power microprocessor first will lower the CPU
P-state. If this action does not achieve the required reduction in
power, it will continue to reduce CPU clock speed to prevent peak
power consumption from exceeding the cap. (HP does not recommend
setting a cap that would throttle performance at a single server
level, blade group or enclosure level.)
When applied at the enclosure level, dynamic power capping
spreads the administrator-defined cap across multiple servers,
dynamically adjusting the power caps for individual server blades,
based on workload intensity. The feature reduces power caps for
server blades running lighter workloads and increases power caps
for server blades running more intense workloads. Because workload
intensity will peak and subside at different times in most
environments, the ability to adjust power caps dynamically lets IT
administrators set the enclosure-level dynamic power cap below the
sum of peak power consumption for each blade without affecting
blade performance.
15
-
16
Protecting service continuity with automated cost-effective
failover and disaster recovery CloudSystem Matrix protects
continuity of services by providing, enabling, or enhancing a range
of availability and recovery solutions. For ProLiant and Integrity
platforms, these solutions range from server-aware and
application-aware availability to disaster recovery solutions for
distances from campus to continental and supporting both physical
and virtual server environments.
Server failover and high availability With BladeSystem at its
core, CloudSystem Matrix has redundant components, so the failure
of a single component such as a power supply or fan does not affect
operations. In addition, you can optionally configure the Central
Management Server (CMS) for Matrix into a high availability
clustered environment, whether a single or federated CMS model.
(For more information see the white paper, Backing up and restoring
HP Insight Software Central Management Server, at
www.hp.com/go/matrixoe/docs). Even if the CMS is unavailable for
some reason, the CloudSystem Matrix environment (and workloads)
will continue to operate in its current configuration.
A simple but powerful feature of CloudSystem Matrix is server
failover: if a server fails, the application can restart on a spare
server and thus be running again in about the time it takes to
power on another server blade, boot the OS (which resides on shared
storage), and start the application service. Failover can take
place within an enclosure, within a data center, or between data
centers. Failover requires no special software to be run on the
servers. It can draw from a pool of spare servers, so the total
number of spare servers can be less than the number of protected
operational servers.
Disaster recovery CloudSystem Matrix recovery management
provides an easy to use disaster recovery solution for ProLiant
servers that protect workloads running on either physical blades or
virtual servers using VMware ESX VM or Microsoft Hyper-V VM
software. The solution is licensed for every ProLiant server blade
in CloudSystem Matrix.
The two-site recovery solution enables push-button disaster
recovery, automating the transition of production workloads from
their preferred site to a remote recovery site in the event of a
site-wide disaster or planned outage. Subsequent failback is
similarly supported. Matrix recovery management can be implemented
across metro, continental, or intercontinental distances.
Matrix recovery management integrates with replicated storage
environments to ensure proper transition of data to the recovery
location and restoration of access to storage from the recovered
servers. Out-of-the-box integration is enabled for HP EVA and HP XP
disk arrays using Continuous Access technology, and for HP 3PAR
storage using synchronous replication. Additionally, administrators
can integrate their own storage failover scripts into Matrix
recovery management using the User Defined Storage Adapter
interface, enabling fully automated, one-touch disaster recovery of
other SAN storage solutions.
Matrix recovery management enables flexible and cost effective
protection against site-wide disasters with support for
physical-to-physical (p2p) and virtual-to-virtual (v2v) server
recovery, as well as cross-technology physical-to-virtual (p2v)
recovery between physical servers and ESX VMs. The
physical-to-virtual recovery scenario can deliver extra cost
savings by enabling consolidation of multiple physical workloads to
a single virtual machine host with multiple guests.
HP CloudSystem Matrix also supports dynamic recovery models,
which enables recovery to either a physical or a VMware virtual
resource, without locking in a recovery target in advance. In this
scenario, the administrator can set up a workload to fail over from
a physical server to either a physical server or a VMware virtual
server, depending on the failover preference specified and resource
availability. If the preferred (physical or virtual) server type is
not available at the time of recovery, the workload can recover to
an available alternative (virtual or physical) server. The
automated recovery of a given workload to either a virtual or a
physical environment can enhance the cost-effectiveness of a
disaster recovery implementation by enabling more flexible use of
the target recovery site.
For additional efficiency and flexibility in recovery
management, Matrix also supports both uni directional and
bi-directional disaster recovery. Uni-directional recovery
configurations are designed to protect services running in a single
production environment. In the event of a site-wide disaster at the
production facility, all disaster recovery-protected workloads can
be failed over to a dedicated recovery site, located remotely,
where production operations can then resume. Once the preferred
production site has been brought back online, workloads can be
failed back from the recovery site to the preferred production
site. The relative simplicity of this architecture can be an
appealing benefit. In situations where production servers need to
operate in different locations, a bi-directional disaster
recovery
-
17
configuration can be used. In such configurations, each site
functions as both a preferred site for some number of servers, as
well as a secondary (recovery) site for additional servers that
normally run at the paired site.
For those environments where business-critical and
mission-critical availability and recovery are required,
CloudSystem Matrix integrates with application-aware solutions to
provide a faster and more granular level of solution recovery. For
Integrity servers, the HP Serviceguard family of products
(available through an upgrade) provides the level of
mission-critical availability that customers require in these
environments. Integrated with key CloudSystem Matrix technologies,
Serviceguard defines packages that define how to start, stop, and
monitor a database or any customer application along with all of
the required resources. The software monitors for faults in
hardware (servers, storage, and networking), in the OS, and in the
application package environment. In the event of a fault, it
automatically initiates a failover sequence to a new node (physical
or virtual) and restores services in seconds, often completely
transparently to application users. Serviceguard manages failovers
within a data center and between one, two, or three data centers by
using HP Metrocluster and Continental cluster technologies.
Application-specific extensions and tools ease deployment and
manageability through the standard configuration of popular
applications in Serviceguard packages. For example, Serviceguard
Extensions for SAP and for Oracle provide tight integration with
SAP and Oracle RAC, delivering superior availability for these
mission-critical environments. (For more information see the white
paper, Wielding HPs Serviceguard Solutions Availability Portfolio
to Maintain Mission-Critical Service-Level Objectives, at
http://h20195.www2.hp.com/V2/GetPDF.aspx/4AA2-3913ENW.pdf, and
listen to the podcast, HP-IT Uses Serviceguard Solutions, at
www.hp.com/go/ServiceguardSolutions.)
Enabling technology HP Converged Infrastructure technologies at
the core of CloudSystem MatrixBladeSystem with Virtual Connect, the
Matrix Operating Environment, and CSA for Matrixenable the powerful
capabilities just described that enable you to auto-provision,
continuously optimize, and protect the continuity of an IaaS
environment. This section provides more information on these and
other, complementary technologies available as extensions to
CloudSystem.
BladeSystem c-Class In designing the HP BladeSystem
architecture, HP worked very closely with our customers to
understand their requirements and challenges in managing their data
centers. The resulting BladeSystem c-Class design incorporates
modular and flexible compute, embedded management, network, and
storage resources to provide a common, modular infrastructure that
is cloud-ready and can accommodate continually changing business
needs. As such, BladeSystem c-Class provides the ideal foundation
for CloudSystem.
Key components of this architecture include the following:
Embedded management: HP iLO Management Engine is a comprehensive
set of embedded management features supporting the complete
lifecycle of the server, from initial deployment, through ongoing
management, to service alerting and remote support.
Provisioning: Intelligent Provisioning takes all the strengths
of products like HP SmartStart, Service Pack for ProLiant, and
Smart Update Manager, enhances them with the latest ease of use
features and places them where you can use them immediately: on the
system board.
Monitoring: HP now offers agentless management. Base hardware
health monitoring and alerting functions now run directly on the
iLO hardware, independent of the operating system and the x86
processor.
Diagnostics: While HP server failures are rare, some failures
can be extremely difficult to reproduce, can escape routine
diagnostics, and consequently may take too long to fix. With HP
Active Health System, diagnostics are always running in the
background, recording constant feeds of telemetry data, every
configuration change, and every alert. This facilitates faster
root-cause analysis and problem resolution, all achieved without
impact to performance.
Support: With iLO Management Engine administrators also get
faster access to HP Insight Remote Support, HPs 24x7 phone home
remote support software.
Shared cooling and power: HP consolidated power and cooling
resources conserve power and provide efficient cooling, while
efficiently sharing and managing the resources within the
enclosure. HP Thermal Logic capabilities throughout the BladeSystem
c-Class servers enable IT administrators to optimize their power
and thermal environments. These capabilities include:
-
18
HP Active Cool fansHigh-performance, high-efficiency, and
hot-pluggable fans provide redundant cooling across each
enclosure
HP Sea of SensorsEmbedded sensors within every blade and
enclosure provide constant information to optimize cooling and save
energy
HP 94% efficiency power suppliesPlatinum level power supplies
offer the highest level of power efficiency available in the IT
industry
HP Power RegulatorEmbedded technology constantly monitors
processor utilization, automatically throttles the processor input
power and frequency to match the application load, and returns the
processors to full capacity whenever necessary
HP Dynamic Power SaverPower mode enables more efficient use of
power in the server blade enclosure by placing power supplies in
standby or active modes based on server utilization levels
Virtualized connectivity: HP Virtual Connect technology provides
a way to virtualize the server I/O connections to Ethernet and
Fibre Channel networks. (See Virtual Connect.)
HP Insight Control power management software leverages embedded
BladeSystem capabilities to provide:
Power RegulatorDynamic or static control of CPU performance and
power states Dynamic Power CappingReclaiming of trapped power and
cooling capacity, as described in Advanced power
management: dynamic power capping
Power MeteringReal-time measurements of actual power consumption
to enable finding and fixing inefficiencies Calibrated Maximum
PowerA power planning tool based on accurate measurement of peak
potential power for
each device instead of estimates
Data Center Thermal MappingGeneration of a thermal map of the
room based on aggregating data from thermal sensors in the
servers
Data Center Power ControlProtection for critical loads by
reducing power to non-critical servers during cooling system
failures
For additional background on HP BladeSystem technologies, see
the BladeSystem technical resources at
http://h18004.www1.hp.com/products/blades/components/c-class-tech-function.html.
For details on servers and OS supported by Insight Control power
management capabilities, see the HP Insight Software Support Matrix
at www.hp.com/go/insightcontrol/docs. For more information on HP
Insight Control software, visit www.hp.com/go/insightcontrol.
Virtual Connect Virtual Connect is an essential foundation
element in CloudSystem. Inherent in CloudSystem functionality and
the powerful automation capabilities of its operations is the
ability to move workloads from one server to another without human
intervention or coordination. Virtual Connect technology enables
this capability.
Most LAN and SAN networks rely on the unique addresses of NICs
and host bus adapters (HBAs) within each server. Replacement of a
server necessitates changes to the MAC addresses and WWNs
associated with network adapters and HBAs on the server and
adjustments in the LANs and SANs attached to those servers. As a
result, even routine server changes are often subject to delays for
coordination among IT operations groups.
Virtual Connect brings all of the necessary capabilities into
the domain of the system administrator. Virtual Connect is a set of
interconnect modules and embedded software that implements
server-edge virtualization. It adds an abstraction layer between
the servers and the external networks. It assigns and holds all MAC
addresses and WWNs at the server bay, instead of on the servers
themselves. At the addition of a new server, its NICs inherit their
assigned server bay MAC addresses and the HBAs inherit their WWNs.
Similarly, upon removal of a server, its replacement inherits the
same addresses so that the LANs and SANs do not see server changes
and do not require updating for them. Using the local Virtual
Connect Manager, or the data-center-wide HP Virtual Connect
Enterprise Manager software, the system administrator can
preprovision and preassign to the enclosure all of the LANs and
SANs that the server pool might ever need (from those predefined by
the network and storage administrators)even before any servers are
installed. SAN and LAN administrators retain sufficient control
over their domains but are freed from interruptions for routine
server maintenance.
-
Through an extension of the same mechanism, Virtual Connect
establishes server connection profiles for each server, and an
administrator can move these profiles from one server bay to
another in a single enclosure or across the data center with a
single mouse click, within seconds. When these capabilities combine
with boot from SAN and the Virtual Connect management tools,
administrators can move application workloads from one server to
another very quickly, securely, and automatically and with their
storage and network configurations intact. The change is
transparent to application users, because the OS image and data
reside on shared storage accessible from a pool of servers. Figure
11 compares HP Virtual Connect to the traditional network
model.
Figure 11: Comparison of HP Virtual Connect to traditional
model
LAN
SAN
NIC
HBA
NIC
HBA
NIC
HBA
NIC
HBA
HP
Virtu
al C
onne
ct M
odul
es
Server SystemServer System
Traditional ModelServer changes impact LAN & SAN
resources
HP Virtual ConnectLAN & SAN resources remain constantas the
server environment changes
LAN
SAN
NIC
HBA
NIC
HBA
NIC
HBA
NIC
HBA
ENET
Sw
itch
FC
Switc
h
LAN
SAN
NIC
HBA
NIC
HBA
NIC
HBA
NIC
HBA
HP
Virtu
al C
onne
ct M
odul
es
LAN
SAN
NIC
HBA
NIC
HBA
NIC
HBA
NIC
HBA
NIC
HBA
NIC
HBA
NIC
HBA
NIC
HBA
HP
Virtu
al C
onne
ct M
odul
esH
P Vi
rtual
Con
nect
Mod
ules
Server SystemServer System
Traditional ModelServer changes impact LAN & SAN
resources
HP Virtual ConnectLAN & SAN resources remain constantas the
server environment changes
LAN
SAN
NIC
HBA
NIC
HBA
NIC
HBA
NIC
HBA
ENET
Sw
itch
FC
Switc
hLAN
SAN
NIC
HBA
NIC
HBA
NIC
HBA
NIC
HBA
NIC
HBA
NIC
HBA
NIC
HBA
NIC
HBA
ENET
Sw
itch
FC
Switc
h
HP FlexFabric, built on Virtual Connect Flex-10 technology and
HP Networking innovation, enables fully virtualized network
connections and capacity from the edge to the core. This enables IT
to deliver network-as-a service, wire connections once, and move
applications freely across or between servers or even across or
between data centers. HP Virtual Connect FlexFabric modules enable
Matrix to connect to any Fibre Channel, Ethernet, and iSCSI network
with a single device. By eliminating the need for multiple
interconnects, Virtual Connect FlexFabric modules reduce network
equipment needs up to 95% and the power needed to drive them by
40%.3 Virtual Connect FlexFabric modules allocate the bandwidth of
a single 10 Gb network port into up to three independent FlexNIC
and one FlexHBA or four independent FlexNIC server connections.
Administrators can dynamically adjust the bandwidth for each
FlexNIC and FlexHBA connection in increments of 100 Mb between 100
Mb and 10 Gb. The result is network capacity applied where
applications need it rather than oversubscribed everywhere.
HP Virtual Connect architecture uses industry standard Ethernet
and Fibre Channel for simple integration with familiar brands, such
as Brocade, Cisco, and HP Networking. The Virtual Connect
architecture is built into every HP BladeSystem enclosure, taking
advantage of the high performance interconnect channels, integrated
I/O connections, embedded communication and control channels, and
modular interconnect bays delivered as standard with HP
BladeSystem. (More information on HP Virtual Connect technology is
available at www.hp.com/go/virtualconnect.)
HP is committed to serving the diverse needs of modern data
centers without imposing a specific operating model, proprietary
architecture, or network fabric. HP is continuing to build out the
Converged Infrastructure vision with the addition of storage
network protocols to its Virtual Connect Flex-10 technology. This
combines the capabilities of Virtual Connect and Flex-10 with Data
Center Bridging (DCB), Fibre Channel over Ethernet (FCoE), and
accelerated iSCSI technologies to enable BladeSystem customers to
use a single pair of Virtual Connect interconnect modules to access
both storage and server networks.
Ultimately, the HP goal is to allow IT to plug new systems into
a Converged Infrastructure that will automatically discover
capacity, add it to resource pools, and put it to work to support
the needs of business applications. For more information on HP
FlexFabric, see www.hp.com/go/flexfabric.
3 Based on HP analysis of networking equipment (adapters and
enclosure interconnects)
19
-
Matrix Operating Environment The HP Matrix Operating Environment
(Matrix OE) is cloud management software for infrastructure as a
service (IaaS) and is at the core of the HP CloudSystem Matrix. It
provides an integrated toolkit enabling the capabilities described
in How CloudSystem Matrix is used:
Design and provision infrastructure services in minutes via a
self-service portal Optimize infrastructure for capacity planning
and showback/chargeback Protect service continuity with automated
cost-effective failover
Matrix OE provides full capabilities for both HP ProLiant and HP
Integrity servers as well as provisioning for third-party x86
virtual machine environments. Matrix OE gives IT staff a full view
and easy access to all functions to manage your Matrix
infrastructure. Figure 12 shows the home page.
Figure 12: Home page for the Matrix Operating Environment
Matrix OE includes and is tightly integrated with HP Insight
Control and enables administrators to manage HP BladeSystem c-Class
blades, standalone physical servers, virtual machine hosts and
guests on ProLiant and Integrity servers as well as physical
hardware partitions on Integrity servers, all from a single
management console. In particular, HP Insight Control delivers
capabilities that enable the proactive management of ProLiant
server healthwhether physical or virtual, the quick deployment of
ProLiant servers, optimized power consumption, and control of
ProLiant servers from anywhere. Matrix OE enables management of
both physical blade servers and virtual machines as logical servers
so they can be moved and migrated easily within a CloudSystem
Matrix environment. It delivers a full range of deployment,
management, capacity planning, migration, and movement
capabilities.
Matrix OE runs on a Central Management Server (CMS). A federated
CMS model permits scaling Matrix infrastructure orchestration
capabilities to manage up to four times the number of resources
possible when using a single CMS. In this model, a primary CMS,
with Matrix infrastructure orchestration software installed, can
direct automated provisioning of resources across multiple
secondary CMSs. The Matrix federated CMS architecture supports
multi-site operationswith secondary CMSs at geographically
distributed locations. For more information on federated CMS
environments, please refer to the current documentation in the HP
Insight Software library at: www.hp.com/go/matrixoe/docs.
Multi-tenancy features in Matrix OE Matrix OE infrastructure
orchestration includes multi-tenancy features for heterogeneous
managed environments, as described in Enabling multi-tenancy. These
environments support VM hosts running on blade or rack-mounted
servers, physical blade servers, private cloud resources, and
clusters. One such feature is support for VLAN network
segmentation, which enhances isolation between Organizations in a
multi-tenant environment. The software uses
20
-
21
VLAN tags in network packets for data separation. This enables a
single wire to carry network packets from separate network
broadcast domains managed by switching technology in the data
center infrastructure and in virtualization hypervisors such as ESX
and Hyper-V. Matrix configures and manages VLAN segmentation with
Operations Orchestration workflows, and uses the Network Automation
utility for network configuration and management. The Service
Provider administrator is able to use these services to only expose
certain segments of the network to each Organization. This allows
for the separation of VLANs between Organizations and secure
network access within each Organization.
Matrix also supports vSphere resource pooling, where a capacity
pool specifies the capacity of the resource instead of VM hosts.
This allows you to specify a pool of CPU and memory belonging to a
host or a cluster, providing better resource sharing. Use of
capacity pools offers additional security for multi-tenancy by
shielding the actual VM location from Organizations, allowing you
to share a cluster or host between different Organizations.
For more information on CloudSystem Matrix multi-tenancy
capabilities see the white paper, Multi-Tenancy in HP Matrix
Operating Environment Infrastructure Orchestration at:
http://h20195.www2.hp.com/V2/GetDocument.aspx?docname=4AA3-9202ENW&cc=us&lc=en.
Cloud Service Automation for Matrix HP Cloud Service Automation
(CSA) for Matrix incorporates a starter edition of HP Server
Automation software and HP SiteScope software. Server Automation
automates lifecycle management for physical and virtual servers and
applicationsfrom establishing a baseline to provisioning, patching,
configuration management, and compliance assurance. SiteScope
provides agentless monitoring of infrastructure platforms and key
performance indicators for applications such as CPU, disk, and
memory usage. As described in How CloudSystem Matrix is used, CSA
for Matrix combines with Matrix OE to enable seamless deployment,
monitoring, and management of applications along with
infrastructure services in HP CloudSystem Matrix.
CSA for Matrix is available as an optional, integrated solution
for CloudSystem Matrix for ProLiant. Comparable functionality is
available for CloudSystem Matrix with HP-UX with the addition of HP
Server Automation and SiteScope software. For more information, see
the HP Cloud Service Automation for Matrix data sheet,
http://h20195.www2.hp.com/V2/GetDocument.aspx?docname=4AA3-1176ENUS&cc=us&lc=en.
Storage technologies Key to the flexibility of the Matrix
solution is having the OS image and data reside on shared storage.
Logical servers can then migrate from one physical server to
another and retain their personality, and continue to provide
application support after reboot. For physical servers, Matrix
supports Fibre Channel storage. Matrix can connect to your existing
SAN or you can order Matrix with an HP 3PAR Storage System or an HP
EVA or XP disk array.
For virtual servers, Matrix supports VMware, Microsoft Hyper-V,
and Integrity Virtual Machines whose boot and data storage are
files within the hypervisor file system. For VMware and Hyper-V,
that file system could be on local disk, FC SAN, or iSCSI. For
iSCSI SAN storage solutions, HP recommends the HP P4300/P4500/P4800
G2 SAN solutions. In VMware environments, Matrix also supports
hypervisor file systems on NAS and virtual machines using Raw
Device Mapping (RDM) to access a FC LUN directly. RDM is a key
enabler to allow a logical server to move from a physical
environment to a virtual one yet still access the same boot and
data volumes. For Integrity Virtual Machine environments, Matrix
supports the use of SLVM (Shared Logical Volume Manager) volumes on
physical FC disks.
For the most current information on supported and recommended
storage solutions, see the HP CloudSystem Matrix Compatibility
Chart at www.hp.com/go/matrixcompatibility.
HP 3PAR Utility Storage is designed to provide the agility and
efficiency specifically required by virtual and cloud data centers.
The integration of HP 3PAR F-Class or T-Class Storage Systems with
CloudSystem Matrix provides you with a simplified way to provision
and scale storage for private cloud environments. The use of thin
provisioning technologies reduces overall capacity requirements and
keeps utilization rates high over time, enabling you to save 50% or
more on the cost of a storage technology refresh.4 For general
information on the full benefits of HP 3PAR storage, see the HP
3PAR Utility Storage Benefits Summary white paper at
http://h20195.www2.hp.com/V2/GetPDF.aspx/4AA3-4024ENW.pdf.
4 Based on deploying HP 3PAR Storage Systems and HP 3PAR Thin
Provisioning and Thin Conversion Software. See Thin Technologies
at
http://h18006.www1.hp.com/storage/solutions/3par/technologies.html.
-
22
The HP Storage Provisioning Manager (SPM), which is included as
part of the Matrix Operating Environment, provides the means for
storage architects to create storage templates, used for all
storage provisioning requests, and for storage administrators to
publish a storage catalog of storage volumes that can be consumed
as needed.
Using SPM the storage architect can define storage templates
containing parameters and policies appropriate for the intended
environment. For instance, template parameters may include volume
size and RAID level along with optional tags. Tags, or naming
conventions, can indicate quality-of-service levelsgold, silver, or
bronze for examplebased on performance and/or availability
characteristics, identify storage volumes according to intended
usagesuch as production or database, or signify other user-defined
criteria. Policies in the templates can control how storage
resources are provisioned such as:
Fulfill with on-demand provisioned volumes (with appropriate
restrictions such as a maximum capacity), pre-provisioned volumes,
or both
Use a particular type of array, a particular array, or a
particular storage pool on an array
Use (or do not use) thin provisioning
Once the storage architect specifies these parameters and
policies within the template definitions, they are applied
repeatedly as the storage templates are used for fulfillment of
service requests. By populating the storage catalog with
pre-provisioned storage volumes (and associated templates) as well
as storage templates for volumes to be automatically provisioned on
demand, the storage administrator is freed from the need to respond
each time a request for storage is made. The storage administrator
retains control of the storage catalog, storage-related processes,
and permissions.
When CloudSystem Matrix is paired with the HP 3PAR Storage
System or the HP EVA solution, the storage catalog in SPM can
automatically discover existing volumes and enable the storage
administrator to import them into the storage catalog and assign
authorizations, thus avoiding manual entry. The storage
administrator can also load volumes created in other arrays (HP or
third-party) into the storage catalog, automatically using
workflows or manually. In addition, the storage catalog can
represent volumes from multiple arrays in the same catalog.
For customers wanting to integrate with their own storage
provisioning processes, whether automated or manual, there is
command line access to the storage definitions (requests) and the
ability to import results back into the Matrix storage definitions.
For more information, see the HP Storage Provisioning Manager User
Guide at www.hp.com/go/matrixoe/docs.
Security The CloudSystem Matrix solution provides a consistent,
integrated, and configurable security management framework,
utilizing security mechanisms in individual components and security
services across the integrated solution.
The Matrix security framework is designed to: Ensure the
confidentiality and integrity of management communications Log all
actions that manipulate server instances Provide separation of
duties and role-based access control for provisioning and
management
Matrix security mechanisms include identification and
authentication, access control, authorization and auditing, and use
of secure communication protocols. To support its security
framework, Matrix makes use of services and facilities such as HP
Systems Insight Manager (HP SIM) core authentication services,
Microsoft Active Directory (if available), the HP SIM audit
facility, the SSL (Secure Socket Layer) transport protocol, the SSH
(Secure Shell) network protocol, and HTTPS (HTTP Secure) for
web-based communications. For a detailed discussion of security in
Matrix, see the white paper on HP Converged Infrastructure solution
security at
http://bizsupport2.austin.hp.com/bc/docs/support/SupportManual/c01755823/c01755823.pdf.
Important extensions available for CloudSystem for security
(described in HP CloudSystem Extensions) include HP TippingPoint
technology, HP ArcSight software, and HP Fortify software.
Integrating HP CloudSystem Matrix into customer environments
Because it is an open, standards-based solution, HP CloudSystem
Matrix works seamlessly with your existing operating systems,
hypervisors, networks, SANs, and applications. In addition, the
Matrix Operating Environment
-
23
integrates with HP and third-party management software and
manages third-party servers to ensure effective management of your
entire IT environment.
Support for most common operating systems and hypervisors Matrix
supports these native server operating systems: Microsoft Windows
Server, HP-UX, Red Hat Enterprise Linux (RHEL), and Novell SUSE
Linux Enterprise Server (SLES).
Matrix supports VMware ESX and Microsoft Windows Server 2008
Hyper-V hypervisors, as well as Integrity Virtual Machines.
For the latest and most complete information on CloudSystem
Matrix compatibility, see www.hp.com/go/matrixcompatibility.
Compatible with existing networks Matrix connects seamlessly to
existing Ethernet networks. Because Virtual Connect is a layer-2
bridge, not a switch, it integrates smoothly with any existing
network. HP Virtual Connect Flex-10 and FlexFabric interoperate
with any industry standard Ethernet switch while providing 4-to-1
network hardware consolidation of the server NICs and interconnect
modules.
Compatible with existing SANs The HP Virtual Connect 8 Gb Fibre
Channel and FlexFabric interconnects are fully compatible with all
standard Fibre Channel switches. The shared storage for logical
server boot and data disks can be any pre-presented FC LUN,
although there are additional capabilities with HP storage
solutions, such as validation of the storage definitions and the
ability to import discovered volumes into the storage catalog.
HP Cloud Maps for accelerated cloud service design HP has
collaborated with key industry-vertical and infrastructure software
partners to deliver Cloud Maps for HP CloudSystem and help you fast
track the design of cloud services tailored to your environment.
Cloud Map pre-packaged workflows and guides can save days or weeks
of solution design time. Cloud Maps contain tested engineering
components, such as best practice templates for hardware and
software configurations, workflows, sizers, and deployment scripts,
along with supporting white papers to guide customization. For
access to the most up-to-date HP Cloud Map resources available,
visit www.hp.com/go/cloudmaps.
HP CloudSystem Developers Resource Center for IT and business
operations integration With growing interest from partners and
customers in HP CloudSystem, HP has seen the need to provide a
single point of access to a greater knowledge base for CloudSystem.
This online resource center presents the CloudSystem information in
an organized way to help accelerate cloud solutions. Here you will
find best practices and links to the most important HP CloudSystem
web pages. Included are tips on taking advantage of the CloudSystem
embedded workflow automation engine and HP Cloud APIs to enable
integration with IT and business processes and initiation of
lifecycle operations on infrastructure services. Examples include
linking into chargeback systems and customizing the self-service
portal to reflect company branding and standards. The resource
center also contains highlighted toolkits and samples content
showing the packages addressing specific integration, automation,
and customization use cases. You can find the HP CloudSystem
Developers Resource Center at www.hp.com/go/csdevelopers.
Management of the extended infrastructure From the CloudSystem
Matrix CMS, you can manage not only the Matrix infrastructure, but
also select ProLiant, Integrity, and third-party servers that the
Matrix Operating Environment manages. Supported capabilities
include automated provisioning of virtual machines, capacity
planning, and recovery management. For details on server models
supported by the Matrix Operating Environment, see the latest HP
Insight Software Support Matrix at
www.hp.com/go/insightcontrol/docs (refer to the Matrix OE supported
servers); for information on licensing requirements, see the Matrix
OE for ProLiant Quick Specs,
http://h18000.www1.hp.com/products/quickspecs/13050_div/13050_div.html.
Integrated virtualization management The IaaS cloud
managementMatrix OEbuilds on years of experience in virtualization
management with HP Integrity systems, combined with HP expertise in
integrated management with key virtualization partners, VMware and
Microsoft. It gives IT administrators a unified method to manage
both VMware ESX/ESXi and Windows Server 2008 Hyper-V virtualization
technologies.
-
24
Matrix infrastructure orchestration supports use of VMware
vSphere native templates to deploy VMware VMs as well as use of
Microsoft System Center Virtual Machine Manager (SCVMM) native
templates to deploy Hyper-V VMs.
Matrix API integration works with tools such as VMware vCenter,
vSphere, and vMotion, and with Microsoft System Center. The
combined HP and partner capabilities provide robust lifecycle
management of virtual environments, including automated deployment,
migration, capacity and power management, health monitoring, and
disaster recovery. For example, when Insight Management receives a
ProLiant pre-failure hardware alert, it works with VMware vCenter
to initiate the movement of all the virtual machines on that server
before the failure occurs.
Matrix OE takes full advantage of the vSphere 5 Auto Deploy
capability to provide automated ESX host and cluster provisioning.
This support allows you to use a cloud infrastructure service
template to create and add a single VM host to an existing cluster
or create a new cluster with a number of VM hosts. The provisioning
process starts from bare metal hardware and is fully automated
through embedded Operations Orchestration workflows to include:
Adding network and storage configuration Installing software
Registering with vCenter server Making the VM host available for
infrastructure orchestration management and VM deployment
Expanding beyond IaaS: CloudSystem portfolio Whether youre an
Enterprise or Service Provider, when you are ready to move beyond
an on-premise IaaS cloud environment, HP has a solution for you on
your cloud journey: HP CloudSystem Enterprise and CloudSystem
Service Provider.
To expand your on-premise cloud infrastructure to a hybrid
private/public cloud model and the full range of service delivery
models (IaaS, PaaS, and SaaS), HP Technology Services can assist
you in making the transition to CloudSystem Enterprise. Like all
CloudSystem offerings, CloudSystem Enterprise is built on the
modular HP BladeSystem architecture and the Matrix Operating
Environment, with the addition of full-featured Cloud Service
Automation software. Cloud Service Automation manages the entire
cloud lifecycle, including advanced application and infrastructure
lifecycle management.
For more information on CloudSystem Enterprise, see the Moving
Beyond IaaS with HP CloudSystem Enterprise whitepaper at:
http://h20195.www2.hp.com/V2/GetDocument.aspx?docname=4AA3-6846ENW&cc=us&lc=en.
HP CloudSystem extensions CloudSystem extensions build on the
core technologies and capabilities to provide an optimized platform
for your cloud infrastructure. Available extensions for CloudSystem
include:
HP TippingPoint technologyOffers continuous protection of both
physical and virtual assets from a single, integrated offering,
allowing you to deploy security policies that automatically adapt
to changes in virtual environments, such as introducing a new
virtual machine.
HP ArcSight softwareCan help organizations safeguard physical
and virtual digital assets, comply with corporate and regulatory
policy, and control the internal and external risks associated with
cybertheft, cyberfraud, cyberwarfare, and cyberespionage.
HP Fortify softwareA suite of integrated applications for
identifying, prioritizing, and fixing security vulnerabilities in
software, dramatically reduces the risk of catastrophic attacks on
applications deployed as cloud applications on the Internet and
helps ensure compliance with government and regulatory
mandates.
HP softwareHP Application Lifecycle Management software
accelerates application development with a test and development
private cloud environment on HP CloudSystem to improve service
quality, end-user performance, and availability management.
HP 3PARCustomers familiar with utility storage can build out an
HP CloudSystem using HP 3PAR technology for the most advanced cloud
infrastructure available. Consider that four out of five service
providers in the Gartner Magic Quadrant Leaders are HP 3PAR
customers.
HP FlexNetworkCloud computing reshapes the way applications are
deployed and consumed and influences data center network designs.
HP CloudSystem together with our HP FlexNetwork solutions can
deliver a unified, open, cloud-optimized data center network design
reducing workload mobility transit time by 80% and doubling VM
application performance.
-
25
For more information on CloudSystem extensions, see the
Understanding the HP CloudSystem Reference Architecture white paper
at http://h18006.www1.hp.com/storage/pdfs/4AA3-4548ENW.pdf.
Purchase and delivery Customers can implement CloudSystem Matrix
by ordering a Starter Kit containing all the necessary hardware and
software or by using Matrix Conversion Services. The Matrix
Conversion Services program allows customers to convert their
existing BladeSystem environment to Matrix without purchasing a
separate Starter Kit.
The Matrix Starter Kit includes all the infrastructure required
for a working Matrix environment:
A fully-redundant c7000 enclosure Virtual Connect modules
Software to enable all of the advanced Matrix functionality for an
enclosure fully loaded with ProLiant or Integrity
server blades
Hardware and software support Implementation services
Customers populate the enclosure with the desired number and
type of server blades and add storage, CMS, rack, and other options
appropriate for different Starter Kit configurations (Table 1).
Table 1. HP CloudSystem Matrix Starter Kits
Matrix with Virtual Connect Flex-10 & FC
Matrix with Virtual Connect FlexFabric
Matrix with HP-UX & Virtual Connect Flex-10 & FC
Matrix with HP-UX & Virtual Connect FlexFabric
Virtual Connect modules
Virtual Connect Flex-10 & Virtual Connect 8 Gb
FC
Virtual Connect FlexFabric
Virtual Connect Flex-10 & Virtual Connect 8 Gb
FC
Virtual Connect FlexFabric
Virtual Connect Enterprise Manager for the enclosure Software
licenses
Matrix Operating Environment for up to 16 ProLiant server
blades
CSA for Matrix (optional)
HP-UX VSE-OE* for up to 4 2-socket Integrity server blades
Server Automation and SiteScope (optional)
Support 3-year, 24x7 4-hour response hardware support
3-year, 24x7 software support and updates**
Blades supported
ProLiant or Integrity Integrity or ProLiant
*HP UX Virtual Server Operating Environment includes the Matrix
Operating Environment with Integrity Virtual Machines and the right
to use the HP UX Operating System as a guest
**Performing upgrade to Matrix environment requires purchase of
a separate service
Customers wanting to minimize their equipment can order one of
the Matrix Starter Kits with Virtual Connect FlexFabric modules.
Doing so enables you to combine Ethernet and storage networks onto
one converged fabric within the BladeSystem enclosure and eliminate
the need for FC HBAs and interconnect modules. For traditional
server edge implementations having separate networks for Ethernet
and storage, the Matrix Starter Kits with Virtual Connect Flex 10
and 8 Gb FC modules are available. In all configurations, Matrix
connects directly to any industry standard Ethernet and Fibre
Channel networks.
The CloudSystem Matrix Starter Kits with HP-UX are optimized for
use with Integrity server blades. Each kit includes eight
per-socket HP-UX 11i v3 Virtual Server Environment-Operating
Environment (VSE-OE) licenses. VSE-OE includes the HP UX Operating
System as well as the Matrix Operating Environment for Integrity
featuring the Integrity Virtual Machines. Similarly, the
CloudSystem Matrix Starter Kits for ProLiant are optimized for use
with ProLiant server blades. These Starter Kits include the Matrix
Operating Environment licenses for 16 ProLiant server blades. HP
supports mixed ProLiant and Integrity server blade environments;
mixed orders require the addition of the appropriate management
software licenses not included with the Starter Kit.
A Windows-based CMS hosts the Matrix management softwarethe
Matrix Operating Environment, consisting of Matrix infrastructure
orchestration, Matrix capacity planning, and Matrix recovery
management for ProLiant, as well
-
26
as Insight Control (with HP SIM), Virtual Connect Enterprise
Manager, and Storage Provisioning Manager. HP offers the ProLiant
DL360 G8 as the recommended option for a CMS. Using this rack-mount
server rather than one of the server blades allows you to use all
available and licensed enclosure blade bays for managing target
blades and workloads. Though the ProLiant DL360 G8 is the
recommended default CMS option, you can substitute any other server
of your choice provided the server matches the Insight Management
CMS requirements (see the HP Insight Software Support Matrix at
www.hp.com/go/matrixoe/docs).
Customers purchasing Matrix Starter and Expansion Kits have the
option of including the HP CSA for Matrix software for important
IaaS capabilities such as basic application deployment and
monitoring and automated patching and compliance. Customers
purchasing Matrix Starter and Expansion Kits with HP UX can add HP
Server Automation and SiteScope for these IaaS capabilities.
CloudSystem Matrix with HP UX allows seamless upgrades of the HP UX
VSE OE to the HP UX Data Center Operating Environment (DC OE).
Upgrading to DC OE adds HP Serviceguard for mission-critical high
availability and disaster recovery.
Fibre Channel storage options from HP include the HP 3PAR F
Class and T Class Storage Systems and HP EVA and XP FC arrays . For
iSCSI SAN storage, HP recommends the HP P4000 G2 SAN solutions.
Matrix also supports third-party FC SANs; for details on the SANs
supported, see www.hp.com/go/matrixcompatibility.
You can scale the capacity of the Matrix infrastructure by
ordering one or more Expansion Kits, either with the initial Matrix
purchase or as needs grow. Table 2 shows recommended small, medium,
and large configurations for CloudSystem Matrix using ProLiant and
Integrity server blades.
Table 2. Recommended configurations for CloudSystem Matrix for
ProLiant and CloudSystem Matrix with HP-UX
SMALL MEDIUM LARGE
ProLiant configurations
Servers Matrix Starter Kit
8 ProLiant Server Blades
DL360 CMS
Matrix Starter Kit
16 ProLiant Server Blades
DL360 CMS
Matrix Starter Kit
3 Expansion Kits
64 ProLiant Server Blades
DL360 CMS
Storage EVA P6300 5.4 TB* 3PAR F200 9.5 TB* 3PAR F400 48 TB*
Software Matrix Operating Environment
CSA for Matrix*
3rd Party Extensions: VMware or MS Hyper-V *
Services CloudSystem Matrix Implement