How to Build a Private Cloud

If you’re nervous about running your business applications on a public cloud, many experts recommend that you opt for a private cloud instead. But building a private cloud within your data center is not just another infrastructure project.

An internal, on-premise private cloud begins with data center consolidation, rationalization of OS, hardware and software platforms, and virtualization up and down the stack servers, storage, and network. Elasticity and pay-as-you-go pricing are guiding principles that allow for the standardization, automation, and commoditization of IT. But it goes beyond infrastructure and provisioning resources. It’s also about application building and the user’s experience with IT. Despite all the hype, internal clouds are still at an early stage. Only 5% of large enterprises are even capable of running an internal cloud, and only about half of those actually do. If you’re interested in being at the forefront of this movement, here’s what you need to know about how to build a private cloud.

First Steps: Standardization, Automation, Shared Resources

To go forward with building a private cloud on-premises, you must have standardized and documented procedures for operating, deploying, and maintaining that cloud environment.

Standardized operating procedures that allow efficiency and consistency are critical for automation, which enables self-service capabilities. In terms of the private cloud, self-service means that an enterprise has established an automated workflow whereby resource requests go through an approvals process.

Once approved, the cloud platform automatically deploys the specified environment. Depending on their needs, developers typically specify the parameters for VMs, storage volume, and bandwidth.

Understand Your Services

Many enterprises misguidedly think about building a private cloud from a product perspective before they consider services and service requirements, but services need to be considered first. Whether a workload has affinity with a private, public, or hybrid model depends on a number of attributes, including obvious ones like compliance and security but also things like latency and interdependencies of components in applications.

If you’re going to commit to building a private cloud, you need to know what your services are, and what the service-level agreements, costs, and road maps are for each of those. Common services with relatively static interfaces, even if your business is highly reliant on them, are those you should be considering for cloud-style computing. E-mail is one example: You may use it frequently, but instead of wanting it to be integrated tightly with the company, you want to make it as separate as possible, easy to use, and available from self-service interface. If you’ve customized this type of service over time, you’ve got to make it as standard as possible. Once you understand which services are right for the cloud and how long it will take to get them to a public-readiness state, you’ll be ready to start to look at building a private cloud from a technology perspective.

Four tiers of components for building a private cloud

First is the resource tier comprising infrastructure, platforms, or software. Raw virtualization comes to mind immediately. Rapid re-provisioning technology is another option. Above the resource pool sits the resource management tier, where the pool is managed in an automated manner.

These two levels are fairly mature. Next comes the service management tier. Here you want something that lets you do service governance and convert pools of resources into service levels. You ultimately need to be able to present to the user some kind of service-level interface that says “performance” or “availability” and have the services management tier delivering on that.

Sitting atop it all is the access management tier, which is all about the user self-service interface. It presents a service catalog, gives users all the knobs to turn, and lets you manage subscribers. The interface has to be tied in some way to costing and chargeback, or at least metering”?at that level, it ties to the service management tier.

It’s all about the business

While building a private cloud means you need to think in terms of elasticity, automation, self-service, and chargeback, you shouldn’t be too rigid about the distinctions at this stage of cloud’s evolution. You might get to SaaS eventually, and in the meantime do as much automation as you can, introducing concepts slowly so your organization has time to adapt to the cloud model. The bottom line is, you have to think about the value the cloud will bring to your organization. Cloud platform providers such as Apprenda have been noted for their ability to make multi-tenancy less complex, to make architecture more mature, and to allow businesses to direct their energy toward efficiency and innovation.