Many organizations are asking whether it makes more sense to move their enterprise to the cloud — or bring cloud computing into their enterprise.
And a surprising number are discovering that the best, most practical answer may be “both.”
What does it mean to bring cloud computing into the data center? And exactly how do you go about creating and managing a data center infrastructure
that provides all the advantages and benefits of cloud computing without needing to leave the enterprise? Basics like virtualization and workload
management are certainly important.
But building private cloud is also about incorporating a “business services” layer of abstraction to your traditional data center infrastructure. In
traditional data centers, computing power and storage capacity create the infrastructure to run server workloads, which are then used to run enterprise
Adding this new business services layer logically groups these server workloads by the business services they support. This shifts the management
focus to full business services — and away from the individual underlying components.
The exact nature of this new business services layer may vary depending on the unique requirements of your organization, but it needs to include a few
core characteristics and capabilities. Here are a few of the non-negotiable functions every service-driven data center should be able to perform:
1. Simplify and accelerate the deployment of business services
Effective service-driven data centers dramatically simplify the provisioning process and reduce the time and effort required to deliver new business services.
This includes automating as many steps of the deployment process as possible, so your infrastructure team spends less time on manual, low-value
processes. Additionally, allowing business service managers and application owners more opportunities to perform a controlled set of limited management
tasks themselves translates directly into higher service levels, more efficiency and fewer distractions for your infrastructure team.
2. Create, publish and deploy standard infrastructure offerings
A service-driven data center also provides a simple model for defining infrastructure offerings and making them available to business service managers.
With the right management tools, you can build a flexible repository of standard infrastructure offerings — including things like standard server
images, pools of storage capacity, standard network access, and so on — and then publish those offerings and make them available for fast, efficient
deployment. This allows business service managers and application owners to browse through a list of available infrastructure services, assemble and
customize the components they need, see the associated costs and then quickly deploy new workloads to support their business service.
3. Add visibility and accountability to current and future capacity needs
Next, your service-driven infrastructure needs to include tools for understanding and managing current and future capacity demands in the private cloud.
Business service managers and application owners tend to treat cloud resources as a free, inexhaustible pool, which leads to ineffective utilization of
resources. Attaching concrete costs to specific cloud resources eliminates this problem and encourages responsible usage and deployment practices.
Your service-driven environment should also help you accurately predict future demands on your private cloud. This includes providing a “pipeline” that
gives business service managers the ability to enter basic information about future capacity needs without actually deploying new workloads. By collecting
all this current and future usage and capacity information in one place and analyzing it carefully, infrastructure managers can gain a deeper understanding of
how cloud resources are being used and how demands will change over time.
4. Keep your infrastructure options open
Growing demand inevitably leads to the need for new servers, storage hardware, operating systems, and other underlying physical infrastructure
components. As you build your private cloud, you should always consider vendor lock-in. Some cloud management solutions are designed to support a
heterogeneous virtualized environment. Consider whether you’ll want the freedom to implement best-of-breed infrastructure components and whether your
group will one day use a hybrid cloud computing model that leverages a diverse range of private and public cloud resources.
5. Provide advanced business service management capabilities
Finally, a successful private cloud environment has to offer relevant insights, deep visibility and efficient management capabilities to everyone with a vested
interest in specific business services, including business service managers, business service owners, application owners and infrastructure teams. This
includes providing the ability to integrate and unify different silos of information; gather, normalize and correlate all the information about your private cloud
infrastructure; and map a wide range of physical, virtual and logical components onto a simple, meaningful service model dashboard. You should also be
able to create customized dashboards for people with different roles, responsibilities and interests.
Making your own internal data center environment more “cloud like” allows you to immediately tap into many of the advantages of cloud computing,
leverage existing infrastructure investments and avoid the current questions and risks associated with public clouds. This incremental approach also makes it
easier to transition to a hybrid public/private cloud model in the future — after security, auditing and other public cloud issues have been addressed.
In other words, many organizations view the creation of private clouds and service-driven data centers as important stepping stones to a more full-blown
private and public cloud computing model.
John Stetic is VP of Product Management at Novell. He is the former Director of Product Management and co-founder of PlateSpin