Step-By-Step Guide to Building a Private Cloud

Deploying a private cloud isn’t just another engineering project. It involves hardware, software, data, infrastructure, budgets and other strategic considerations.

By Dipti Parmar

By Dipti Parmar April 19, 2022

It might not be obvious, but even in the face of the seemingly endless growth and capabilities of public cloud systems, the private cloud is holding its own. In fact, a quarter of all respondents to our latest Enterprise Cloud Index study maintained that their IT environment ran on the private cloud – the second most popular deployment model after multicloud.

Further, cloud repatriation (reverse migration) from the public cloud is a now a thing – up to 85% companies have considered moving their applications or workloads back to on-prem or private cloud environments at some point.

Needless to say, the decision to deploy a private cloud should be a well-thought out and strategic one. Enterprises need to define and lay out their needs, expectations, and goals clearly before attempting to build a private cloud model. Here’s an attempt to list out the various considerations and steps that go into this endeavor.

Making the choice to go private

Cloud means something different to every organization, depending on its size, industry, and the nature of the data it deals with. Further, there is the private vs public vs hybrid cloud debate (which the hybrid cloud appears to be winning).

Choosing private cloud over public mainly boils down to two things: control (in terms of security and privacy) and budgeting (choosing CAPEX over OPEX).

The private cloud is often a default choice for companies operating in the finance, health, and telecom spheres, where customer data falls under the purview of government and industry regulations.

“There is only so much governance and corporate pain you can wrap around a line of business,” said Ed Hoppitt, a veteran in the cloud app and platform development space. “So first understand why it’s private cloud. If it’s regulatory or security issues, the focus then becomes very much on building the right security model and policies, making sure they can still operate in a way that reflects the kind of agility someone would expect from public cloud.”

To make an informed decision, the organization must be clear about its cloud computing as well as overall technology goals.

Define the requirements and clarify outcomes

Most organizations make the leap to private cloud after they’ve built up core IT strengths in virtualization, consolidated their data center or hardware, and succeeded in optimizing IT infrastructure costs to a great extent.

The next step for them is advanced virtualization, standardizing operating procedures, automating more processes (to a deeper level), and enabling self-service capabilities for IT resources.

In a typical private cloud-enabled workflow automation scenario, developers request VMs with specific memory, storage, and bandwidth parameters, the resource requests through a quick approval and provisioning process, and then the platform automatically deploys the approved environment.

Organizations migrate to or adopt the private cloud in order to meet their upgradation, performance, security, or compliance goals, but they still want cost-effective capacity management without any bottlenecks.

This requires a careful assessment of existing resources as well as accurate estimation of outcomes. Right sizing of hardware resources and auto-provisioning of correctly configured VMs is no mean task on a cloud scale.

“If a VM isn’t being used, shut it down – as long as there are no processes going behind to clean up,” said Tiago Fernandes, Cloud Solutions Architect at Tech Data. “I don’t see why you can’t do snoozing in private cloud as well.”

And that’s why it would be a mistake to think about building a private cloud from a product perspective before considering services, application dependencies, and workload affinity. According to the Nutanix Enterprise Cloud Index report, an equal proportion – around 35% – of all enterprise workloads run on private and public clouds (the rest are deployed on traditional data centers) but more teams chose the private cloud for database, disaster recovery, HR, ERP, and big data applications.

Obviously, IT needs to know every service inside out, draft service level agreements, work out costs and draw roadmaps for each workload before committing to the private cloud.

It’s up to the cloud architect to identify and scope out dormant resources. They also need to know about upcoming projects or strategies that might affect demand, so an unhindered information flow to and from customer service, marketing, purchase, and pretty much every department is necessary.

“Capacity management affects ROI and everybody in the business,” said Fernandes. “Everybody looks at other departments and says: you didn’t forecast. However, IT needs to be connected with the business to be able to predict future demand.”

Capacity planning, therefore, is key to delivering optimized resources to users and getting a viable level of ROI out of the private cloud. Constant monitoring of cloud metrics and periodic stress-testing of applications and workloads should be part of the cloud management strategy.

Build on the HCI advantage

A hyperconverged infrastructure (HCI) is the ideal foundation for hosting private clouds. It combines universal data center hardware elements with intelligent, purpose-built software to provide the ideal platform for a cloud environment.

A HCI is an intermediate step between legacy IT infrastructures and the private cloud. It “converges” on-premise assets such as servers, storage networks, and storage arrays (such as NAS or SAN) into a streamlined infrastructure, leaving the door open for upgradation into a hybrid IT infrastructure or hybrid/multi-cloud environment.

This makes building a private cloud easy because a complete data center stack – including compute, storage, and networking components – is topped off with virtualization and combined into industry-standard, commodity “nodes” that can scale up one at a time. Purpose-built software runs on each node to distribute all operating functions – for every workload – across all of them. Collectively, this set of nodes forms a “a cluster” that delivers outstanding performance and superior resilience.

HCI offers a clear separation of application process logic (or business rules), storage components, and data access that enhances interoperability between the private cloud and other cloud systems as and when the need arises.

This is made possible by the two primary constituents of HCI:

  • The distributed plane, which runs on a cluster to deliver server virtualization, storage, and networking services to guest applications that run on VMs or containers
  • The management plane, which provides a single, unified administration interface for all cloud and on-prem resources

Some other architectural features of HCI that support and sustain private cloud deployment are:

  • Full control over security configurations and audits
  • Better data protection, data-at-rest encryption, and simplified data management
  • Micro-segmentation of workloads
  • Built-in backup and disaster recovery
  • Fast and automated deployment
  • IT as a Service (ITaaS) (which enables automation, self-service, and integration across clouds)
  • Reduced operating expenses
  • Container based services
  • Enhanced application mobility
  • Cloud-native application development

Deploy in steps

The actual deployment, while technically the most complex part, is achievable if the IT team has planned it well with the approval of leadership.

Setting up a private cloud infrastructure involves three fundamental steps:

  1. Set up compute, network, and storage resources with clusters. Typically, a private cloud starts out with at least two machines or clusters that can be loaded with all the resources a VM needs.
  2. Install management software for the hardware. Usually, the software is specific to the stack but Nutanix installations come with a combination of hardware and software, so admins can skip this step.
  3. Choose and configure a backup solution (on a per-VM or full-cloud basis) and set up servers for redundancy.
  4. Configure private and public network addresses, and NAT if the workload specifications require it.
  5. Define admin roles and add users. Set up security policies and authentication methods.
  6. Install applications, provision VMs, and create storage containers. Create VM templates and configure licensing where necessary.
  7. Publish application blueprints and make them available to developers for self-service provisioning.

Cloud next

Demand for the private cloud is still growing. An IDC survey with over 2,200 respondents found that private cloud will comprise 40% of all cloud deployments by 2023. Moreover, 44% of enterprises plan to increase their private cloud spending this year.

A trend that’s hastening private cloud adoption is cloud MSPs and “package” cloud vendors offering “private cloud as a service” – pre-built, pre-configured, pre-secured systems (delivered in HCI-based racks) that can scale up quickly. These are charged on a pay-as-you-grow basis and reduce operational costs even more.

While there’s no denying that hybrid, multicloud environments are taking precedence, workloads in the private cloud are no longer limited to “what can’t go on public clouds.” Organizations are increasingly realizing the benefits of more control over security, data handling, and infrastructure governance.

Dipti Parmar is a marketing consultant and contributing writer to Nutanix. She’s a columnist for major tech and business publications such as IDG’s CIO.com, Adobe’s CMO.com, Entrepreneur Mag, and Inc. Follow Dipti on Twitter @dipTparmar or connect with her on LinkedIn for little specks of gold-dust-insights.

© 2022 Nutanix, Inc. All rights reserved.  For additional legal information, please go here.