Cloud Service Models
Offloading Responsibility.
One of the simplest explanations of cloud services is the idea of offloading responsibility. In a traditional client-server model, companies do not offload responsibility for their servers, storage, or networking. Instead, they purchase expensive, complex, and often proprietary hardware, hire a team of experts to manage it, and cannot easily scale their investment as their business needs change. With the cloud, businesses can offload some or all of those responsibilities to a cloud service provider (CSP).
Cloud service models are initially divided into three types of solutions, differentiated by what responsibilities are being offloaded. Many more cloud service models have since been defined, but these five best exemplify the primary aspects of cloud computing: SaaS, PaaS, IaaS, FaaS, and XaaS.
Read carefully! This lesson contains critical information and terms used extensively in the course.
Software as a Service.
Software as a Service (SaaS) permits consumers to use the software provided by the cloud service provider (CSP). The CSP retains responsibility for installing, configuring, maintaining, patching, and upgrading the software. The software is typically accessible from many client device platforms, such as phones, tablets, and traditional computers.
Licensing is usually based on a subscription model, where only the number of deployments needed is purchased. SaaS is also operating system agnostic—meaning organizations rarely have to worry about an employee's preferred OS. For example, an organization could deploy a SaaS solution like Dropbox that supports Windows, macOS, and Linux seamlessly, whereas a traditional non-cloud implementation would struggle to provide a centralized, policy-managed solution across all three.
Platform as a Service.
Platform as a Service (PaaS) scenarios provide consumers with the hardware, operating system, and necessary tools. The consumers then utilize the tools to manage their data on their own. The CSP is responsible for hardware support, operating system support, and platform maintenance. The consumer simply uses the platform within the scope of their business needs.
PaaS solutions are often aimed at developers and database administrators (DBAs). By offloading responsibility for the hardware and OS to the CSP, developers do not have to concern themselves with supporting the platform they are working on. Menial tasks, such as OS and application updates, hardware failures, and device drivers, are no longer a concern. A DBA no longer has to request a server, install an OS, and install a database platform like Microsoft SQL Server or MariaDB—those tasks are offloaded to the CSP.
Infrastructure as a Service.
Infrastructure as a Service (IaaS) offloads responsibility for hardware support to the CSP. Consumers are responsible for managing virtual machines hosted on the CSP hardware infrastructure. The consumers install operating systems (OS) themselves, such as Microsoft Windows or Red Hat Enterprise Linux (RHEL), configure and patch the OSs, and install software.
IaaS can be much more cost-effective because consumers can deploy the exact server platforms they need. In a traditional infrastructure model, companies must predict server needs during a budget cycle and estimate overall utilization. The pay-as-you-go model of IaaS means organizations can pay for what they need—no more and no less. By offloading physical space, power, cooling, and security requirements to the CSP, companies may save a great deal of money and time in Operating Expenditures (OpEx).
Function as a Service.
Function as a Service (FaaS) enables developers to easily create code that executes in response to specific events without creating and managing the underlying compute infrastructure. FaaS is also called serverless for its independence from the infrastructure. Developers may use FaaS to host microservices, including data processing, real-time analysis, or Internet of Things (IoT) functions.
Serverless computing still utilizes compute resources, contrary to what the name implies. Compute resources are allocated on demand to applications, and no resources are reserved when the application is not in use. Billing reflects the application’s actual use of resources. Serverless environments require no configuration, monitoring, or capacity planning.
Anything as a Service.
Anything as a Service (XaaS), where “X” represents any possible service, began as a way to take the agility, speed, and reduced initial capital expenditures that occur with the other cloud-based “as a service” deployments and shift those benefits to other IT services. These services might include email, desktop operating systems, remote access, and even security.
Many innovative companies have creatively carved a niche within the “as a service” role for themselves. XaaS means any delivery with a flexible, pay-as-you-go structure. Many organizations have embraced this business model for its reduced capital expenditures and business agility.