What is a virtual private cloud (VPC)?

Private cloud is a cloud computing environment where all hardware and software resources are dedicated and exclusively accessed by a single customer, either through the internet or a private network connection. It’s also known as ‘internal cloud’ or ‘corporate cloud’. Private cloud can operate either from an on-premise data center, third-party colocation facility, or private cloud provider that offers a private cloud hosting service.

Why use a private cloud

Many organizations choose private cloud for the following advantages:

Security

As private cloud is limited to be exclusively accessed by a single customer, it helps meet security and regulatory compliance requirements such as GDPR. Private cloud also helps to meet the requirements when handling confidential and sensitive data such as personally identifiable information, medical records, intellectual property, financial data, etc.

Sustainable performance

Since private cloud is dedicated to a single customer, the performance is predictable. Public cloud is multi-tenant, meaning noisy neighbors can affect sustainable performance by taking away shared resources.

Cost

The cost of a private cloud is consistent each month, as long as you are using the same hardware and software. If you own the hardware and facility, private cloud could yield long-term savings with predictable cost compared to monthly fees paid for public cloud.

Control

Private cloud gives you the freedom to choose whichever hardware and software in any configuration that makes sense to you. Since you have full control over the hardware in a private cloud, you will get greater visibility and access. Conversely,   with public cloud, you will have to choose from existing offerings by cloud service providers (CSPs).

How private clouds work

Private clouds may use solutions from Microsoft, VMWare, Red Hat, OpenStack, and others. They can also be deployed in a single region, single location, or multi-region and multiple locations; it can aggregate resources from these locations as pools of resources, then, based on need, divide into smaller virtual resources by using following technologies:

• Virtualization: Private clouds provide virtualization technologies to abstract physical hardwares, such as computing, storage, memory into pools of resources. These resources can be partitioned into multiple virtual machines, containers etc. This allows resources to be shared efficiently amongst multiple applications and users, and provides the flexibility to dynamically create and remove virtual machines to provide agility and scalability of the cloud.
• Software-defined networking (SDN): Physical data centers require you to add/remove cabling and log in to hardware devices to enable and configure specific ports and routing. Private clouds use SDN that abstracts the underlying networking layer, and provides dynamic, programmatic methods for creating virtual networks.
• Single management panel: Instead of managing individual resources and treating them as pets, a single, centralized management panel gives you control of the entire infrastructure to have a birds-eye view of utilizations, availability, and security of the private cloud environment. 
• Automation and DevOps: Configuring individual resources manually is time-consuming and error-prone. With automation, you can reduce human error and time to delivery for the creation of virtual resources, and you may further integrate automation into the CICD pipeline.

Differences between private clouds and public or hybrid clouds

Public CSPs  (such as Amazon Web Services , Microsoft Azure, Google Cloud Platform, and Oracle Cloud) own multi-regional, multi-location data center facilities and provide multi-tenant services to customers, accessible via the internet or private connections. The most common CSP pricing model is pay-as-you-go, which is great for workloads that don’t run 24/7 and provides lower cost of entry. For long-running workloads, CSPs may provide a reserved pricing model that locks into monthly or yearly terms to help reduce cost. Since the environment is multi-tenant, public cloud customers share the infrastructure with each other, but advertised performance is normally peak performance and not sustainable for a long period of time. On the plus side, CSPs are constantly innovating their products in the public cloud, which gives immediate access to cutting-edge technologies. CSPs also have data centers that are much higher capacity than private data centers, meaning CSPs can offer greater elasticity and scalability.

Hybrid clouds are the combination of private cloud and public cloud infrastructure, using the same management panel. Customers can choose or move workloads between private cloud and public cloud freely to meet their applications’ security, compliance, performance and cost requirements.

As an example, hybrid cloud allows you to extend your bursting workload from your private cloud portion to your public cloud portion, where security and compliance isn’t an issue. You may also choose to move workloads from your private to public cloud portion if you need additional temporary computing/storage and don’t want to pay for fixed cost.

Private cloud deployment models

On-premises private clouds

Private cloud is deployed in an organization’s own data center. The organization has full responsibility for hardware and hosting facilities and thus must take care of maintenance, operation, security, etc. It’s ideal for organizations that have the right resources to manage the private cloud.

Hosted private clouds

Private cloud is deployed in a hosted data center outside of the organization’s facility. The organization must provide hardware, but hosting facilities (racks, cooling, fire suppression systems, electricity, physical access control, etc) are provided by the hosting provider. It’s ideal for organizations that want to use a private cloud, but do not want to operate their own data center.

Managed private clouds

Private cloud is deployed in either an organization’s own data center or in a hosting provider’s facility. The organization still must provide hardware, but will outsource the operation and maintenance tasks in lieu of a dedicated operations team.

Private cloud providers

• Hewlett Packard Enterprise
• VMWare
• Microsoft Azure Stack via Dell (hybrid cloud)
• Oracle Private Cloud Appliance
• IBM private cloud with Red Hat
• Cisco Private Cloud
• NetApp
• Amazon Web Services

Pros and cons of private clouds

Pros of private clouds

• Security: All hardware/software are exclusively accessible by a single customer. This helps organizations maintain strict compliance and security requirements, while still enjoying the agility and elasticity of a public cloud. 
• Guaranteed performance: No one else is sharing the hardware, and all computing/storage/networking are dedicated to a single organization, which gives predictable performance.
• Cost: While public clouds have low entry pricing due to the PAYG model, long-running workloads will enjoy long-term cost savings in private clouds.
• Control: Private cloud gives organizations full control. Not only do enterprises get to choose their own hardware, software, networking, etc, but they also get lower level visibility of packets, cabling, etc.

Cons of private clouds

• Management: in both on-premises and hosted private cloud deployment models, organizations need to have dedicated operations teams to take care of the hardware and private cloud operations. if the private cloud operation is in hundreds of nodes, then this can be outsourced.
• Cost: Hardware has to be purchased up front, which results in capital expenses. Without proper price-to-performance planning, the long-term cost saving may not be realized.
• Capacity and scalability: The organization is responsible for capacity planning and actively needs to review and remove resources when no longer needed, which adds additional operational overhead. Furthermore, additional resources may need to be purchased ahead of time in case of some spiking workload, as private cloud may not have enough capacity to complete the workload on time.