Multi-tenancy is the practice of having one running software instance, or one shared pool of computing infrastructure, serve many separate customers at the same time. Each customer is called a “tenant,” and the defining requirement is that one tenant’s data and activity are isolated from every other tenant’s, even though they share the same underlying system.
The idea is closely tied to the NIST definition of cloud computing (NIST SP 800-145, September 2011), which lists “resource pooling” as one of the five essential characteristics of the cloud. NIST describes the provider’s resources as being pooled “to serve multiple consumers using a multi-tenant model, with different physical and virtual resources dynamically assigned and reassigned according to consumer demand,” typically with the customer having no control or knowledge of the exact physical location of those resources.
Multi-tenancy is the economic engine behind cloud and software-as-a-service pricing. Because the fixed costs of servers, software, and operations are spread across many tenants rather than dedicated to one, the per-customer cost falls dramatically. This is what lets a SaaS vendor offer the same application to thousands of organizations affordably, and what lets a cloud provider keep utilization high.
The trade-off is that isolation must be engineered carefully. Tenants share memory, storage, and processing, so the system has to guarantee that no tenant can see or affect another’s data. Approaches range from a shared application with a tenant identifier on every record, to separate databases or separate virtual machines per tenant, each balancing cost against the strength of isolation.