This primer introduces multitenancy and commodity hardware and explains why they are used by cloud platforms.
Cloud platforms are optimized for cost-efficiency. This optimization is partially driven by the high utilization of services running on cost-efficient hardware that manifests as multitenant services running on commodity hardware.
The decisions made in building the cloud platform also influence the applications that run on it. The impact to the application architecture of cloud-native applications manifests through horizontal scaling and handling failure.
Multitenancy means there are multiple tenants sharing a system. Usually the system is a software application operated by one company, the host, for use by other companies, the tenants. Each tenant company has individual employees who access the software. All employees of a tenant company can be connected within the application while other tenants are invisible; this creates the illusion for each tenant that they are the only customers using the software.
In a single tenant model, if an application needs a database, it gets its own instance. This simplifies capacity management (for individual applications), but at the expense of overall efficiency, as many database servers (and other types of servers) will be running with low overall utilization much of the time.
In the cloud, multitenant services are standard: data services, DNS services, hardware for virtual ...