What Is Public Cloud Computing?
Public cloud is a deployment model that draws on the compute resources of a cloud service provider (CSP). The resources are delivered over the internet, as a service. Public cloud deployments are typically managed on a pay-per-use basis that incurs operating expenses (OpEx) but does not require an investment in hardware as a capital expense (CapEx.) The user organization contracts with the CSP to gain access to the servers and supporting technologies that are owned and maintained by the CSP.
How Public Cloud Works
The public cloud delivers compute resources over the internet from a CSP, such as Amazon Web Services (AWS), Google Cloud, Microsoft Azure, and many other global or regional companies.
The public cloud provides a variety of computing resources in a shared environment. Each user organization, based on its needs, purchases a specific type, level, and amount of performance and bandwidth from the CSP.
In this multitenant architecture, many users may consume the same hardware resources. Individual user organizations may choose to deploy their cloud services on software-defined virtual machines or in containers to help isolate their applications and data within the CSP’s infrastructure and to restrict access to authorized users and customers only.
Migrating workloads to the public cloud requires a disciplined approach across the entire organization. Cloud computing operations are most effective when designed and implemented by a multidisciplinary DevOps team that integrates development with IT operations.
Benefits of the Public Cloud
Public cloud services—in one form or another—have become a must-have in businesses of all sizes. Cloud-based applications, developer platforms, and streaming services are often delivered via public cloud resources, offering many flexible options for computing deployment and consumption.
Because they’re so accessible and easy to scale, public cloud services offer speed and agility to accelerate time to market and the flexibility to experiment with new development tools, applications, and services.
The public cloud model is popular to a large degree because its up-front costs are minimal and require little or no capital investment. Services are typically priced on a per-use basis, making it relatively easy to adjust capacity and performance levels as the organization’s needs change.
Further, public cloud service providers own and maintain all the server hardware, reducing the burden on a user organization’s IT team. Small companies and startups can often launch products and services in the public cloud with little to no in-house IT support.
The public cloud is especially well suited for workloads that run intermittently or temporarily. For example, a startup can adopt public cloud services to scale computing, networking, and storage resources to meet its rapidly evolving needs without incurring the expense and risk of a capital equipment purchase.
Public Cloud Architecture
Public cloud service providers offer a variety of configurations and options, from “bare metal” servers to fully containerized, virtualized computing environments that may include cloud-based storage and networking choices.
The first step in designing a cloud architecture is to assess the organization’s computing environment, workloads, and needs.
CSP instances and offerings can be complex. When migrating workloads to the cloud, organizations can and should take advantage of cloud management tools that help to optimize performance and reduce costs. These tools can match the needs of individual organizations and workloads to the right services, for improved ROI and more-effective cloud deployments.
Public Cloud Security
Some organizations are hesitant to migrate workloads and data to the public cloud because of security or privacy concerns. Individual CSPs may offer encryption and other services to accommodate some cloud security needs. However, some organizations may prefer to continue running their most sensitive applications and data in a private cloud environment for better control.
In some regions, countries, or industries, regulations may restrict the storage location or usage parameters for sensitive data. When regulatory compliance is a critical issue, a private cloud might be the preferred choice.
Public Cloud vs. Private Cloud vs. Hybrid Cloud
The public cloud is one type of cloud deployment model. While a private cloud is powered by infrastructure you own, the public cloud delivers resources over the internet from a CSP.
Some workloads and usage models are best suited to the public cloud, while others are a better fit for a private cloud infrastructure. Many organizations will see the best results with a hybrid cloud approach that combines private and public cloud deployments.
For example, cloud-native organizations and applications might maintain a presence in the public cloud while returning or repatriating certain workloads and data to an on-premises, self-hosted private cloud. The resulting hybrid cloud model is one possible path to a successful cloud computing journey.
Cloud implementation strategies can also move in the opposite direction, from an on-site private cloud infrastructure to the public cloud environment. The migration from private cloud to public cloud can offer opportunities to modernize the underlying infrastructure and shed legacy hardware without additional hardware investments or service interruptions. The private-to-public cloud migration can also lead to a hybrid cloud implementation, either as an interim stage or as an end result.
Public Cloud Examples and Use Cases
Thanks to the wide range of options offered by cloud service providers, public cloud service models can be found or customized to support almost any use case. For example, an Infrastructure as a Service (IaaS) model offers access to hardware resources, but the user organization’s IT team manages the operating system, databases, and data. An IaaS model gives the user organization more control, including opportunities to create virtualized environments with their own virtual resources.
Another option is the Platform as a Service (PaaS) model, where the CSP hosts the hardware, firmware, operating system, and tools, and the user organization takes responsibility only for its own applications and data.
End user organizations face an even lower barrier to entry when adopting the Software as a Service (SaaS) model in the public cloud. SaaS is commonly used to deliver business applications, such as customer relationship management (CRM) or payroll. The user organization engages with the SaaS provider whose software is typically delivered through a web browser. The customer is responsible only for the data.
There are other types of public cloud services as well. Each arrangement can be tailored to the individual needs of the user organization and the sophistication level of its IT team. The common thread among public cloud services is that the user organization can take advantage of an array of computing services without purchasing and maintaining the server hardware.
The public cloud can be especially well suited to artificial intelligence (AI) applications and workloads. AI projects often begin with experimentation and many iterations before they are ready to provide any long-term value to the organization. The trial-and-error processes can include multiple adjustments to hardware configurations, so it would be risky to purchase servers that might not support future needs of the AI solution. Deploying in the public cloud helps organizations to postpone or avoid investing in capital equipment before the solution design is finalized.
Intel® Solutions for the Public Cloud
The public cloud infrastructure worldwide is built on servers, and a large proportion of that server hardware is built on Intel® technology. Over the years, Intel has worked with major cloud service providers to optimize performance on the Intel® technology in their data centers.
Intel also works with ecosystem partners, such as VMware, Red Hat, and Microsoft—as well as top CSPs such as AWS, Google Cloud, and Microsoft Azure—to offer almost endless combinations of public and private cloud resources.
Intel delivers a strong foundation for cloud computing, both on-premises and in the cloud. Since the beginning of cloud computing more than a decade ago, Intel has collaborated with CSPs and others in the ecosystem to deliver coengineered solutions and optimized performance.
Millions of Intel® CPUs are deployed worldwide to support powerful, flexible, scalable cloud computing services. Cloud developers benefit, as they are able to move workloads seamlessly from public to private to hybrid cloud or even multicloud implementations—all running on Intel® hardware—with minimal retesting or revalidation required.
Public Cloud Developer Resources
Large public cloud service providers offer developer frameworks and Intel® integrated development environment (Intel® IDE) tool suites to help developers work faster and more efficiently. Intel also works with CSPs to deliver integrated frameworks and libraries optimized for the cloud services available in the CSPs’ online marketplace.
When choosing a CSP, evaluate their specialized service offerings and match them to the needs of your business to drive faster innovation. For example, many cloud service providers today provide Analytics as a Service, including support for data warehouses and data lakes.
Intel customers and partners can take advantage of our cloud development sandbox, Intel® DevCloud, for testing a variety of Intel® architecture-based configurations that may also be available in CSP instances. Intel and our technology partners offer tools and technologies for cloud management and optimization.
Moving Forward with the Public Cloud
Each organization and application can have unique requirements. A cloud migration strategy should be responsive to those needs with careful choices for workload placement and optimizations. It is also important to continue monitoring costs and balancing them against value across the entire organization.