Public, private, hybrid and multi-cloud

Why deployment models exist

Service models (IaaS, PaaS, SaaS) say how much you manage. Deployment models answer a different question: where does the infrastructure live, and who else shares it? The answer depends on cost, control, law and history. A five-person startup and a hundred-year-old bank make very different choices.

Public cloud

A public cloud is run by a provider and shared by many customers over the internet. Each customer's resources are kept separate (this separation is called isolation), but the hardware pool is shared. You pay as you go and get the most services and the fastest start. AWS, Microsoft Azure and Google Cloud are the three biggest. Others include Oracle Cloud Infrastructure, IBM Cloud and Alibaba Cloud. India also has local providers and data-centre companies, for example Yotta, CtrlS, Sify and Tata Communications, often chosen for government and data-residency work.

Private cloud

A private cloud is used by one organisation only. It can sit in its own data centre or be hosted by a partner. It still offers self-service and automation (for example with VMware or OpenStack software), but the company buys and runs the hardware. You get the most control, but you pay up front and need a skilled team. It is not "a server room with a new name": without self-service and automation, it is just on-premises.

Hybrid cloud

Hybrid cloud joins on-premises or private systems with a public cloud so they work together, usually over a private network link. Banks are the classic example: the core banking system, which holds every account balance, often stays on-premises for control, regulation and history, while the mobile app front end, analytics and test environments run in the public cloud. Providers sell tools for this, such as AWS Outposts (AWS hardware in your building), Azure Arc and Google Distributed Cloud.

Multi-cloud

Multi-cloud means using two or more public cloud providers, for example AWS for the main app and Google Cloud for data analytics. Reasons include using the best service of each, avoiding dependence on one company, a customer's requirement, or a merger that brought in a second cloud. The cost is complexity: two sets of skills, two bills, two security models.

Community and government clouds

A community cloud is shared by organisations with the same needs, such as government departments. In India, the Ministry of Electronics and Information Technology (MeitY) empanels cloud service offerings after an audit, so that government departments can buy cloud services that meet its security and data-location rules. This programme grew out of the GI Cloud initiative called MeghRaj. A related idea is the sovereign cloud: a cloud where data, operations and staff access stay inside one country under that country's laws. India's Digital Personal Data Protection (DPDP) Act 2023 and sector rules from regulators such as the RBI make data location an important question.