The Arrival of the Extended Data Center:

Cloud and On-Premise as an Integrated

Execution Fabric

Abstract

While Cloud Computing offers multiple opportunities for right-sizing the data center, taking advantage of the

Cloud requires changes to IT skills, support processes, and toolsets. And while the Cloud is still viewed as

disruptive technology by some, ENTERPRISE MANAGEMENT ASSOCIATES® (EMA™) analysts see an

ideal end state as seamless integration of Cloud, on-premise, and internal

Instead of over-provisioning to ensure that execution environments are sized for utilization spikes, IaaS

enables companies to provision on-premise infrastructure for usual and expected loads. When they require

additional horsepower for year-end reporting or “number crunching” engineering applications, the Cloud

stands ready to absorb processing overflow.

IaaS vendors, such as Amazon with its Elastic Compute Cloud (EC2), provide raw computing power. They

offer a standard platform that is basically a blank canvas – provisioning, maintaining, and securing IaaS-based

services are the responsibility of the customer.

Among the early adopters of IaaS are development and testing organizations within IT, and the future role of

public clouds in supporting business-critical applications is still uncertain. Regardless, utilization of IaaS-based

services will likely continue to increase over time and, in some cases, under the radar screen of IT executive

management.

Cloud and On-Premise as an Integrated Fabric

While many in the industry still see clear lines of demarcation between “the Cloud” and on-premise, those

lines are blurring. Some leading-edge companies are deploying private clouds, essentially consisting of pooled

resources that are shared across the company. Others are already seeing IaaS as an extension of the data

center, but only for non-core services or periods of exceptional resource utilization.

PRODUCT BRIEF

- 1 -

©2009 Enterprise Management Associates

Cloud. Achieving this vision requires a higher degree of automation and more

sophisticated tools than has been the case in the past.

Phurnace Deliver™, an application deployment and configuration automation

tool, can help companies move toward this integrated approach. Phurnace

Deliver offers a deployment management alternative that automates both

Cloud and on-premise application provisioning, while eliminating manual

scripting and unwieldy runbooks. This EMA Product Brief elaborates on

these ideas and discusses Phurnace Deliver in more detail.

The Cloud as an Execution Platform

The beauty of Cloud Computing lies in the fact that it offers unprecedented

opportunities for “right-sizing” the data center, with virtually unlimited

computing resources available on-demand. Cloud services come in multiple

varieties, including Software as a Service (SaaS), Platform as a Service (PaaS),

and Infrastructure as a Service (IaaS). These variations are evolving rapidly, with

the lines between the three somewhat hazy and often overlapping. Of the three,

IaaS is probably the least understood and, up until now, the least utilized.

Phurnace Deliver offers

a deployment

management

alternative that

automates both Cloud

and on-premise

application

provisioning, while

eliminating manual

scripting and unwieldy

runbooks.

The Arrival of the Extended Data Center:

Cloud and On-Premise as an Integrated

Execution Fabric

While EMA sees the lines between Cloud and on-premise continuing to diminish, IaaS is a special case with

its own unique requirements. Taking advantage of its benefits has implications across the IT enterprise.

Architecting and deploying IaaS services require IT specialists who understand the unique requirements of

on-premise and Cloud-based service delivery. Securing and managing IaaS

Phurnace Deliver: Eliminate Scripts with “Snapshot Application

Provisioning”

For most companies, the complexities of maintaining the application execution fabric are masked by an army

of IT specialists. CIOs lament the fact that the cost of IT administration and support – simply “keeping the

lights on” – continues to rise. However, the time and effort required to knit together and maintain potentially

hundreds of separate execution platforms into a cohesive application fabric are expensive, accounting in part

for the high cost of support.

Automation can help, and one outcome of the recent financial meltdown is that CIOs are turning to

automation as a way to streamline processes and optimize costs. In fact, EMA research finds that

“investigating ways that automation can add efficiency to support processes” is the number one way that the

economic slowdown has impacted budget priorities.1

With its flagship Phurnace Deliver solution, Phurnace Software offers a provisioning alternative that

fundamentally changes the way applications are configured and deployed. Navigator-based and wizard-driven,

it automates traditional, labor-intensive processes by observing application configurations via a specified IP

address. From this observation point, Phurnace leverages the APIs of middleware providers (IBM

WebSphere, Oracle WebLogic, RedHat JBoss, etc.) to collect application-relevant configurations.

Phurnace software discovers interrelationships between applications, their environments, and topologies by

discovering configuration attributes and elements through “expert system” target interrogations. It takes a

“snapshot” of an existing deployment (also described as a “last known good” version), complete with tiered

services, configurations, and software/hardware dependencies (see Figure 1). It then stores the harvested

PRODUCT BRIEF

- 2 -

©2009 Enterprise Management Associates

deployments requires security expertise and tools capable of spanning on-

premise, Cloud, and mixed execution environments. Ideally, at some point

IaaS will become simply an extension of the execution fabric. Today,

however, there are multiple roadblocks thwarting this vision, not the least of

which is provisioning applications, middleware, and operating systems.

Provisioning tiered services, databases, and application servers is tedious,

time-consuming, and error-prone. It requires

careful

analysis and

documentation of application configurations at each tier, often across

multiple operating systems, database platforms, and application servers.

Driven by hand-written scripts and runbooks, simply preparing a deployment

package can require weeks of manual labor. And although the premise behind

Cloud services is instant availability, this potential is wasted if it takes a month

to provision those services.

While EMA sees the

lines between Cloud

and on-premise

continuing to diminish,

IaaS is a special case

with its own unique

requirements.

1

“IT Speaks: Practices and Trends in Enterprise Application Management”, EMA, 2008