Cloud computing - Cloud Service and Deployment Models

Cloud Computing Service and Deployment Models Cloud computing delivers computing resources through different service models and deployment approaches. Understanding these distinctions is crucial because they determine what responsibility the customer bears versus the cloud provider. Service Models Overview Cloud services are typically organized into layers. The key principle across all service models is who manages what: as you move up the stack from Infrastructure to Software, the provider takes on more responsibility, and the customer takes on less. Infrastructure as a Service (IaaS) IaaS provides virtualized computing resources over the internet. Think of it as renting raw computing power. What the provider manages: The underlying hardware, hypervisors (the software layer that creates virtual machines), physical servers, storage, and networking infrastructure. What the customer manages: Operating systems, applications, and data. You install and run whatever software you need. How it works: IaaS uses two primary technologies for resource isolation: Virtual machines run on a hypervisor, which divides a physical server into multiple independent machines Containers use Linux kernel features (control groups and namespaces) to isolate applications more efficiently than virtual machines Cost model: Utility-based billing. You pay for what you allocate and use—similar to how you pay for electricity. This is fundamental to IaaS appeal: no large upfront capital investment. Real-world example: Amazon EC2 or Microsoft Azure Virtual Machines, where you rent servers on which you can install Windows, Linux, databases, and applications. Platform as a Service (PaaS) PaaS provides a development and deployment environment. Instead of managing servers, you focus on building and deploying your application. What the provider manages: Servers, operating systems, storage, networking, and the development platform itself (runtime, libraries, databases). What the customer manages: Only the application code and configuration settings specific to your use case. How it works: The provider offers a ready-to-use environment with pre-installed programming languages, frameworks, and services. You write your code and deploy it; the provider handles scaling, updates, and infrastructure. Variants: PaaS comes in specialized forms: Integration Platform as a Service (iPaaS) focuses on connecting different business systems through integration flows Data Platform as a Service emphasizes data visualization and analytics tools Real-world example: Heroku or Google App Engine, where you deploy code in a supported language and the platform automatically scales it. Software as a Service (SaaS) SaaS delivers fully managed, ready-to-use applications. You don't manage anything technical. What the provider manages: Absolutely everything—infrastructure, platform, application, updates, security patches, and backups. What the customer manages: Only user-specific settings and configurations within the application. How it works: Applications run on the provider's servers and are accessed through thin clients (like web browsers). Multiple organizations share the same application instance (multitenancy), which reduces costs for both the provider and customers. Cost model: Subscription-based pricing, typically per user per month. Automatic updates mean you always have the latest version without doing anything. Real-world example: Salesforce, Microsoft 365, or Gmail—software you use through your browser without installing anything locally. Serverless Computing Serverless is classified as its own service category alongside IaaS, PaaS, and SaaS, though it builds on platform concepts. What it is: A model where customers run code or process data without provisioning or managing any hardware or operating system resources. You upload a function or script, and the platform runs it when triggered by events. Key difference from PaaS: Even more abstraction—you don't manage servers or platforms, only your code. The infrastructure scales automatically based on demand. Real-world example: AWS Lambda or Azure Functions, where you upload a function that runs only when called. Comparing Service Models The image above shows the critical distinction: each service model shifts responsibility. On-premises means you manage everything. As you move right toward SaaS, the provider manages progressively more layers, and you manage less. This is the "responsibility matrix"—a key concept for understanding cloud services. Key insight: The higher the level of service (IaaS → PaaS → SaaS), the less flexibility you have, but the less operational burden you carry. Deployment Models Overview While service models describe what kind of computing resources you get, deployment models describe where and how those resources are hosted. Private Cloud Definition: Cloud infrastructure operated solely for a single organization, managed either internally by the organization or by a third party, but the resources are dedicated to that one organization. Key characteristics: Can be hosted internally (on-premises) or externally (hosted by another company in their data center) Only your organization accesses the resources More control and security, but higher costs Cost implications: Requires significant capital investment in hardware, physical space, and environmental controls (cooling, power). Additionally, hardware must be periodically refreshed as it ages. When to use: When you need strict data control, must meet regulatory requirements (like HIPAA for healthcare), or want to leverage existing infrastructure investments. Public Cloud Definition: Cloud services delivered over the public internet, shared among multiple customers, often offered via subscription or free of charge. Key characteristics: Multi-tenant environment—many organizations share the same infrastructure Accessed over the public internet Provider handles all management and scaling Lower cost per user because costs are distributed Additional capability: Direct-connection services allow you to securely link your legacy data centers to cloud applications without going over the public internet. When to use: When cost efficiency is important, you need rapid scaling, and data sensitivity is not extreme. Hybrid Cloud Definition: A combination of public cloud services with private cloud or on-premises resources. Workloads can move between environments based on needs. Key benefit—cloud bursting: During normal operations, you run applications on your private infrastructure. When demand spikes, extra workload automatically "bursts" into the public cloud. You pay for public cloud capacity only when you actually use it. When to use: When you want flexibility, cost savings during peak periods, and the ability to keep sensitive data private while using public cloud for less sensitive tasks. <extrainfo> Community Cloud Definition: Shared cloud infrastructure among several organizations with common concerns such as security compliance, regulatory requirements, or geographic jurisdiction. Managed internally or by a third party. Cost advantage: Costs are distributed among participating members, offering partial savings compared to maintaining a completely private cloud, while offering more control than a public cloud. When to use: When organizations need to meet strict compliance standards together (e.g., healthcare providers, government agencies). </extrainfo> Multi-Cloud and Poly-Cloud <extrainfo> Multi-cloud: Using public cloud services from two or more cloud service providers. The primary goal is often to increase flexibility and avoid vendor lock-in—not being dependent on a single vendor. Poly-cloud: A variant where you leverage specific services from multiple providers based on what each does best, without the primary goal of redundancy. You choose Provider A's database service because it's superior, Provider B's machine learning services, etc. </extrainfo> Key Takeaway Service models define what you get and who manages it (IaaS, PaaS, SaaS, Serverless). Deployment models define where it's hosted (Private, Public, Hybrid, Community, Multi-Cloud). Together, they give you flexibility to choose the right computing approach for your organization's needs and constraints.