A Guide To Cloud Computing

GlassHouse Technologies White paper

Preparing Your Enterprise for IT-Managed Service Utilities

Jason Goodman,
Technical Writer There’s been a thunderstorm of white noise surrounding cloud computing this year. Vendors,
GlassHouse pundits, and analysts took centre stage at industry events and in the news; everyone seems
to have an opinion and definition on cloud computing. And with all the chatter came mixed
messages, conflicting advice and, well, cloudiness.

There is good news: it’s time to filter out the hype and confusion. While still evolving, cloud
computing is here to stay. Organisations need to be ready.

The fundamental promise of cloud is transformational: shifting from ad hoc, capital-

intensive, IT delivery methods to agile, pay-per-use service utilities, delivered over the
network. A few forward-thinking enterprises have been working on internal clouds. In the
process of consolidating data centres, reducing servers and storage n-fold, and squeezing
every pound out of purchased IT infrastructure, these companies have developed
sophisticated cost analysis, tracking, and service delivery and support techniques. In effect,
becoming less like traditional IT and more like service providers.

But for most organisations, cloud computing is relatively amorphous. For practical
consideration, cloud needs to be addressed in the context of immediate issues and future
opportunities for enterprise IT.

One key challenge for IT executives is developing an enterprise-wide strategy for cloud
computing that’s proactive and drives the business. Creating and communicating this plan
requires IT leadership, knowledge, and guidance around cloud initiatives.
This paper reviews these challenges and benefits in five main sections:

• Overview on cloud computing: from GlassHouse’s perspective

• Cloud computing trends: analysis, delivery and deployment models
• Enterprise cloud readiness: guidelines for planning and evaluating enterprise cloud
solutions, starting with private clouds—and internal cost assessments
• Service provider transformation: requirements and goals shifting from traditional IT
to a service utility, like a service provider
• Cloud Infrastructure Services: an introduction to GlassHouse’s service offerings to
help you get there

© Copyright 2009 GlassHouse Technologies (UK) Ltd. All rights reserved. 1

 Cloud Computing Overview
GlassHouse defines cloud computing as an IT service utility that gives IT the ability to
deliver agile services, at the right cost and the right service level. The network (public or
private) provides an ideal delivery medium for IT service utilities. The deployment model
(internal, external, or some combination) follows the business: your applications, data sets,
security policies and customers.

In general, cloud services should provide the following key characteristics:

• Shared or ‘virtual dedicated’ computing infrastructures to deliver utility computing
services
• On-demand, fluid scalability of all computing resources—server, network, and
storage
• Pay-as-you-go cost models based on actual usage time of consumed resources
• Automated, self-provisioning of services and resources, through a self-service portal
• Ubiquitous network access: the service can be acquired over the network
• Location-independence: the service can be acquired and accessed locally or remotely,
independent of geography
• Integrated monitoring, metering, reporting, billing and chargeback systems
• Predictable cost assessment and management models

From an IT service delivery perspective, internal clouds leverage existing investments to

provide demand forecasting, cost transparency, and automation. Maximising the use of
your existing assets helps reduce costs and improve efficiency.

With true demand forecasting and planning capabilities, you can evolve from a just-in-case
scenario to a just-enough provisioning system for real-time infrastructure allocation.
Cost transparency allows you to develop meaningful business-based metrics to make
smarter, informed decisions for all aspects of your infrastructure—from facilities to staff. In
addition, cost transparency provides the information to enable you to evaluate your capital
versus operational costs and margins, and report on them.

Finally, the automation aspect of an internal cloud provides a key building block for shifting
from an up-front capital to a modular, pay-as-you-go operations model.

© Copyright 2009 GlassHouse Technologies (UK) Ltd. All rights reserved. 2

 Cloud Computing Trends
There is significant innovation happening today in cloud computing and IT should
understand the benefits and drawbacks of various trends, models, offerings—and their
potential impact to the business. GlassHouse believes that decisions made today will impact
an organisation immediately, and influence your roadmap over the next two to five years.

In a recent GlassHouse survey of enterprise IT executives, about 60-percent of executives

plan on implementing cloud initiatives within the year; the other 40-percent have no plans
yet. If your company is in the early stages of investigating cloud computing, you are not
alone: nearly 60-percent of all executives said they had not yet appointed an owner for
cloud computing initiatives.

Wherever you are in the planning phase, 2010 will be a major year for cloud computing.
Gartner estimates the current market in 2009 for cloud services to be nearly $50 billion, and
predicts this will grow to $150 billion in 2013.1 This represents an entire spectrum of cloud
computing providers and ecosystem solutions.

IT leaders can begin assessing and evaluating cloud computing trends by understanding:
• Potential benefits and applicability to the organisation
• Delivery models to compare provider offerings
• Deployment models in support of your organisation’s public and private clouds

Prospective Organisation Profiles

What types of companies will benefit from cloud computing? The short answer is that
companies of all sizes and profiles will benefit, but there are a few key traits of firms that
most benefit. These include but are not limited to:
• Innovative firms investing in research and development that need to remain focused
on business and streamline IT in support of business objectives
• Businesses with rapidly expanding (or contracting) capacity requirements, requiring
on-demand, highly scalable capacity. Growth could be seasonal or punctuated by
unpredictable spikes in usage
• Businesses needing to reduce capital expenses and the practices of funding over-
provisioned hardware resources, on a per-project basis
• Businesses familiar with or requiring Service Level Agreements (SLAs)
• Any businesses needing to develop accurate demand-forecasting capabilities
o Evolving from “just-in-case” to a “just enough” provisioning approach
o Ending the practice of over-provisioning resources

1Source: Gartner Group. Forecast: Sizing the Cloud; Understanding the Opportunities in Cloud
Services. 18 March, 2009. Pg 1.

© Copyright 2009 GlassHouse Technologies (UK) Ltd. All rights reserved. 3

 Delivery Models
Cloud computing presents numerous technologies, terminologies, methodologies, issues,
and deployment models. There are three main categories of offerings: Infrastructure as a
Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS).

Infrastructure as a Service
Examples include many public cloud services such as Amazon EC2 and various service
providers (Rackspace, Savvis, Go Daddy, for example) shifting from pure-play co-location
and hosting to “virtual dedicated” and to shared, multi-tenant cloud services.
About 42-percent of IT executives surveyed by GlassHouse consider Infrastructure as a
Service as the cornerstone of their cloud computing strategy.

Platform as a Service
Major platform vendors (VMware, Microsoft, Google Apps, Force.com) are now calling their
products cloud platforms (or Platform as a Service). Any cloud-enabled application
development platform—often built on Web Services—can be thought of as PaaS. This is an
amorphous category with many variations among vendor solutions.

Software as a Service
Salesforce.com’s success has popularised SaaS (or Software as a Service) as an alterative to
dedicated large chunks of IT infrastructure to customer applications. SaaS can be thought of
as an application or service delivered directly as a service over the network, usually the
internet or intranet.

Deployment models
Deployment models are used to describe privacy (public and private), location, and
ownership (internal, external, and hybrid).

Private vs. Public Clouds

A central decision behind deploying cloud computing is the level of privacy required—and
the profile of users serviced by the cloud.
• Private cloud: a private cloud is exposed only to IT users within an organisation. A
private cloud is a secure cloud and associated set of security services (authentication,
administration, identity management) available to a specified user-base
• Public cloud: a public cloud can mean many things. It could be an internal public
area or collaboration zone (like an extranet) where some aspect of the internal
network could be exposed to users outside the organisation. The term can also refer
to providers offering shared access to shared infrastructure resources, across a public
network

© Copyright 2009 GlassHouse Technologies (UK) Ltd. All rights reserved. 4

 Internal Clouds
Among IT executives surveyed by GlassHouse, 72-percent consider internal clouds their
highest priority or first cloud computing initiative.

An internal cloud is an inward-facing IT service utility, managed locally by IT, designed to

streamline hardware and software infrastructure operations and management. Internal
clouds may be geared around a specific workload or service. The fundamental premise is
that they are private, managed locally, and usually owned by the IT organisation.

While internal clouds are typically managed by IT, some innovative service providers such
as Surgient are beginning to offer on-premise, managed private clouds.

External Clouds
An external cloud is deployed and managed on remote IT infrastructure, either directly by IT
or a 3rd-party vendor. External clouds come in various forms. Large, shared, multi-tenant
clouds such as Amazon and Google provide one example of a public external cloud. In this
case, the service provider provides all resources—hardware, software, and in some cases
management automation—remotely. For companies with minimal regulation, compliance,
and security requirements, this type of external cloud offers no-to-low up-front costs, true
pay-as-you-go usage models, and fluid scalability of resources.

Some service providers (RackSpace and Savvis, for example) offer virtual dedicated,
remotely hosted private clouds, where all computing hardware can be assigned to an
individual customer—but is charged for on a monthly or per-use basis. For customers with
higher security, performance, or scalability requirements, this may be an option.

© Copyright 2009 GlassHouse Technologies (UK) Ltd. All rights reserved. 5

“Cloud readiness requires
viewing current offerings
through the lens of a
service provider. Cloud
vendors offer services with
certain defined
commitments and
associated costs for
delivery. If you cannot
express existing service
capabilities in the same
manner, how can a
meaningful build vs. buy
cost comparison be done?”
-- Jim Damoulakis, CTO of
GlassHouse
© Copyright 2009 GlassHouse Technologies (UK) Ltd. All rights reserved.
Hybrid (Federated) Cloud
A combination of internal and external models is known as a hybrid or federated cloud.
Platform vendors like VMware and Microsoft position hybrid or federated clouds as the
building blocks for cloud computing, where the core evolution of IT functions gradually
shifts from purely internal IT services to incorporating aspects of external or public clouds.
As IT services evolve, aspects of an internal cloud could potentially shift or be outsourced to
remote sites via VPNs and/or other techniques.
Scenarios where hybrid clouds could be used today include but are not limited to: virtual
labs for test, development, and staging; line-of-business application migration or hosting;
and disaster recovery (DR), among others. `
Hybrid clouds will likely grow in popularity among medium- to large-size IT organisations
as more customers deploy virtualisation technology on platforms such as VMware vCloud,
Microsoft Azure, and Citrix Cloud Center (C3).
+) but the means to obtain or integrate cloud services requires a framework and
methodology for gathering and managing detailed cost information in support of services.
Cost Transparency
Cost transparency provides the foundation for tracking costs and showing business value.
To assess the capital and operational savings, companies need to gain visibility into their
current cost structure—to ultimately compare against any new service offering, in-house or
externally procured. Cost transparency provides the basis for developing granular metrics
around cost of capital, cost of operations, and profit margin.
These costs include but may not be limited to:
• Hardware Infrastructure: networks, servers, storage, and components
• Software Infrastructure: software and application licenses, tools, and utilities, etc
• Operations Infrastructure: staff and team required to develop, test, deploy, operate,
and support services
• Facilities Infrastructure: floor space, power and cooling, etc.
Companies may not need to immediately start charging directly for services, but IT can
provide significant cost transparency—and hence chargeback—back to management and the
user community. The costs should be associated with defined services and service levels so
that clients and customers understand what they are getting and where the costs come from.
As your organisation matures, companies can refine and adapt services to internal clients.
The reality is, IT may not have the time, resources, or knowledge to assess, build, and
deploy internal cloud strategies and services—much less analyse, plan, and integrate new
services with current IT infrastructure functions. GlassHouse provides a Cloud Impact
Analysis, described later in the paper, which provides this framework tailored to your
business.
6

“In the long run, an IT
organisation will have to
function as a service
provider. This is normally
a major shift in IT
governance, changing a
significant number of
policies, processes, and
technologies within the
organisation.”
--Rob Zylowski, Services
Director, GlassHouse
© Copyright 2009 GlassHouse Technologies (UK) Ltd. All rights reserved.
Many IT organisation do not have the time, resources, or knowledge to assess, build, and
deploy cloud strategies and services—much less analyse, plan, and integrate cloud services
with existing IT infrastructure.
In support of IT, GlassHouse offers a set of
Cloud Infrastructure Services designed
specifically for IT customers needing to:
• Determine the impact and opportunity
to leverage cloud-based services
• Develop an internal cloud IT services
delivery capability
• Investigate the appropriateness and
applicability of external service
providers
• Optionally build and extend internal
(often private) “hybrid” clouds to utilise
federated, external resources, as aligned with business requirements.
What separates GlassHouse from other IT consulting groups?
GlassHouse provides:
• An objective vendor-independent approach to helping your IT organisation define
and execute the best cloud strategy for your environment, leveraging multi-vendor
solutions already deployed on-premise
• Proven methodologies and expertise in cost analysis, monitoring, and metering to
provide visibility and control for global IT enterprises
• IT transformation services to drive high-performance and agility through cloud
computing—reducing costs, improving efficiency, and managing service levels
effectively.
Application and Data Portability
A central tenant of the external and hybrid deployment models is application portability.
Files (usually virtual machines, databases, or user data files) must be replicated in bulk, or
partially, as changes are made in one location. This poses challenges for proprietary external
clouds (and virtualisation formats) such as Amazon and Google. Once an application is in a
proprietary format, contains large amounts of data and is located in a remote data centre, it
will not easily be replicated across a WAN without significant cost.
This will require you to assess both storage replication and new file transfer (import/export)
technologies to replace or compliment FTP and HTTP, as well as ways to leverage industry-
standard virtualisation software to maximise data portability.
7
 Security and Compliance
Shifting from a dedicated to a shared hosting infrastructure model—internally or
externally—carries security implications. Each deployment model poses different
challenges and consequences. Each cloud provider warrants deeper scrutiny, possibly
through security or accounting audits, to assess vulnerabilities. Fundamentally, the
deployment model should align with your security and business goals.

Business unit customers with high security and compliance requirements may require
dedicating virtual private clouds. These are islands of hardware, networks, and software
instances dedicated to a business unit and locked in a physical jail or cage. Legal
departments, financial services companies, health care organisations and other highly
confidential information may require these higher levels of security.

IT Transformation: Cloud Service Delivery

While there is much focus on the public cloud, where the actual delivery of a cloud service is
the responsibility of a service provider, many organisations are of a sufficient scale that they
will prefer to build and offer internally-owned cloud services first—often as a private cloud
within the organisation. The deployment and management of a private cloud also demands
that IT adopt a similar mentality to a public cloud provider.

IT transformation capabilities required include:

• Specific service offerings defined and made available in a Service Catalogue to
advertise and track SLAs and various Key Performance Indicators (KPIs).
• Co-terminus service level agreements (SLAs) to guarantee various service levels.
• Demand-forecasting in support of real-time or rapid provisioning.
• Billing and chargeback automation.

The end-product of cloud computing within an enterprise IT organisation is the combination

of actual services offered, guaranteed, operated, supported, and billed as well as the users’
ability to easily consume and pay for the services.

© Copyright 2009 GlassHouse Technologies (UK) Ltd. All rights reserved. 8

 Cloud Service Provider (Sample Functions)
The enterprise cloud service provider must be able to perform the following functions:
• Provide a Service Catalogue of standard services and their units of consumption. The
Service Catalogue provides a list and definition of services offered
• Provide stable reliable services with consistent service levels
• Report on service levels and Key Performance Indicators (KPIs)
• Provide the ability for customers to provision and self-manage services
• Ensure that sufficient capacity is available to meet needs and accommodate usage
spikes (or dips) in demand
• Provide non-disruptive, rapid scalability and provisioning of additional
infrastructure, as required
• Provide accurate reporting and usage tracking of consumed resources for billing and
chargeback purposes
• Develop metrics such as cost-per-unit for various cloud services
• Ensure appropriate levels of security are maintained for all services
• Have process or reporting systems to demonstrate cost transparency information—
i.e., cost of services obtained by clients

Cloud Service Consumer (Sample Functions)

An enterprise cloud service customer (consumer) must be able to:
• Understand the services being offered and how they align to end-users’ business
requirements
• Manage request and approval processes for additional resources to control
consumption
• Ensure maximum service utilisation; meaning, provisioned services are fully utilised,
and idle resources are released in a timely manner back to IT to avoid unnecessary
charges
• Understand resource owners and billing so that cloud service costs can be charged
back to individual business consumers or business units
• Understand if internal or outsourced services are providing appropriate levels of
service and the most cost effective solution to the business

Service Level Agreements

The end-product to customers usually requires an SLA. SLAs can be written to support
variable or fixed service levels around capabilities such as performance, availability, data
integrity, and data recovery. For internal clouds, an IT organisation may or may not have
complete control over its SLAs, depending on the nature of the business. For SLAs to be

© Copyright 2009 GlassHouse Technologies (UK) Ltd. All rights reserved. 9

 enforced, a granular set of services and process will likely need to be tracked and clearly
defined.

Automation
Many service providers (from carriers to ISPs to external cloud providers) supporting large
numbers of users, are likely to have rich automation processes in place. These processes are
typically designed for automated provisioning, management, reporting, operations support,
billing, and charge back. If your organisation is not already employing automation, you
should consider applying some of the techniques and methodologies of the service
providers to your IT operations.

© Copyright 2009 GlassHouse Technologies (UK) Ltd. All rights reserved. 10

 GlassHouse Cloud Infrastructure Services
GlassHouse helps customers in all phases of the IT cloud services continuum: from a
strategic offering determining how to best implement a cloud in the customer’s unique
environment to integrating internal and external clouds. GlassHouse offers foundation
services supported by their expertise in virtualisation, security, data protection, and their
tools for monitoring and metering IT environments.
GlassHouse is an IT cloud services enabler providing:
• Services to help customers prepare their environment to work effectively with cloud
technologies, usually starting with an internal cloud strategy
• Strategic direction, design and plan execution for customers establishing IT-specific
private/internal clouds, or shifting aspects of an environment to an external cloud,
often as hybrid or subset of an internal cloud strategy
• Cost assessment and transparency services to help customers assess, monitor and
meter, and visualise their cloud environments—internal and external
• Management automation: through fine-grained monitoring, metering, and
management infrastructure, provided through their Optics Suite, customers gain
deeper visibility and control of their IT environment and cloud deployment

Strategic Services
GlassHouse provides a Cloud Impact Analysis service to determine the strategy, business
requirements, and compelling rationale for moving forward with an internal or external
cloud (private or public). GlassHouse works closely with customers to build a business case
and cost analysis based on ROI modeling for potential cloud solutions. GlassHouse also
recommends solutions and provides a prioritised action plan for moving forward.
For customers requiring modernisation and automation, GlassHouse offers a Service Provider
Model Modernisation service. This is a transformational service that creates the foundation to
operate infrastructure through standardised levels of service and cost. It includes the tools
and processes necessary to automate and operate as a service provider. The package can be
tailored to specific needs, exploring the potential for Cloud services (private or public) for an
entire infrastructure area or discrete services within a larger portfolio. This model lays the
foundation for demand-planning and forecasting processes and cost transparency.

Infrastructure Integration Services

Once a strategy and project plan have been put in place, GlassHouse offers a set of
implementation services for designing, integrating, deploying, and managing cloud services
as IT service utilities.

Accelerate Cloud—an Internal Cloud Jumpstart

Accelerate Cloud focuses on designing, planning, integrating, and deploying cloud
technology through IT. Often, this leverages core building block technologies, including
virtualisation, storage, networking, and other infrastructure components. Accelerate enables
IT to rapidly go-to-market, reduce risk, and increase time to value—at a fixed cost.
Accelerate Cloud will provide clients the technical solution details associated with the latest
cloud Platform as a Service products, such as VMware vSphere.

© Copyright 2009 GlassHouse Technologies (UK) Ltd. All rights reserved. 11

 Managed Services

Optics Suite (Reporting & Monitoring)

Optics is a subscription based suite of services to monitor and report on your infrastructure
components, including storage, backup, databases, virtualised environments and security.
Optics provides deep visibility into IT infrastructure processes, and operations, giving you
control over your environment and demonstrating the business value of the deployed
solution.

Conclusion
While there has been much hype around cloud computing, it is critical for executive staff of
competitive enterprises to understand and begin planning for internal cloud services.
IT executives must own and drive enterprise-wide cloud computing strategies. The
cornerstone of an enterprise cloud computing strategy is based on obtaining accurate
information and guidance, especially for internal cloud readiness.

For enterprise IT to fully realise the efficiency and promise of cloud computing, IT must
demonstrate cost transparency to drive a service-oriented strategy, evaluate prospective
external vendors, and deliver end-customer-consumable cloud services.

GlassHouse provides objective, vendor-independent cloud enablement services to help your

IT organisation determine and deploy the right cloud technologies and management
practices, at the right cost. Through proven methodologies and expertise, GlassHouse
enables IT to deliver internal, external, and hybrid clouds to reduce costs, improve
efficiency, and manage service levels effectively.

To learn more about GlassHouse’s Cloud Infrastructure Services, please email

sales@glasshouse.com or visit the GlassHouse web site to find an office near you.

About GlassHouse

GlassHouse Technologies is a global provider of data centre infrastructure consulting

services. Focused on data centre consolidations, virtualisation, security, storage, data
protection and managed services, GlassHouse consultants offer a vendor-independent
approach to architect, implement and operate IT environments that drive high performance
and agility through cloud computing and a service provider model. GlassHouse transforms
infrastructures to deliver high value services that align to business needs, enabling cloud
computing and accelerated return on investments while mitigating operational inefficiencies
and risk. This is provided through Transom, a unique delivery framework comprised of
proprietary software tools, methodologies and domain expertise. Visit the GlassHouse blog
for expert commentary on key data centre issues facing today’s enterprises and follow us on
twitter at #GlassHouse_Tech.

© Copyright 2009 GlassHouse Technologies (UK) Ltd. All rights reserved. 12