[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"glossary-on-premises::en":3,"gloss-cluster-on-premises::en":23,"gloss-next-on-premises::en":9},{"slug":4,"category":5,"name":6,"definition":7,"meta_desc":8,"faq":9,"schema_markup":9,"related":10},"on-premises","cloud","On-Premises","On-premises means software running on hardware the organisation owns or leases in a facility it controls, rather than on a provider's shared infrastructure. It is worth separating from two things it gets confused with. A private cloud is a pool of virtualised capacity that may well sit in somebody else's building; a single-tenant hosted deployment is a dedicated instance that is still operated by the vendor. On-premises is narrower than both: the machines are yours to buy, house, power, replace and eventually retire. The reasons buyers still ask for it are real. Some data is not permitted to leave a building or a jurisdiction. Some systems it must talk to will never be reachable from outside the network. Some organisations have already bought a decade of hardware and would rather use it. And in a few cases the arithmetic genuinely favours owning capacity you run continuously rather than renting it by the hour. What changes is the shape of the work rather than its amount. Capacity stops being a setting and becomes a purchase with a lead time, which means growth has to be forecast rather than absorbed, and a burst you cannot serve is a queue rather than a bill. Elasticity mostly disappears: the machines are there whether you use them or not, so the incentive flips from spending less to using more of what you already have. Upgrades become events with a change window and a rollback plan instead of something a provider does quietly. And the failure domains you inherit — a power feed, a cooling unit, a single rack — are yours to design around. None of that makes the choice wrong. It makes it a commitment to run an operations function, and that function should be staffed before the hardware arrives rather than after the first outage. A useful test at the proposal stage is to ask who replaces a failed disk on a Sunday and how long the spare takes to arrive, because the answer to that question is the whole of the difference between owning capacity and renting it.","On-premises runs software on hardware you own and house. How it differs from private cloud, and what changes in capacity, elasticity and upgrades.",null,[11,14,17,20],{"slug":12,"name":13},"disaster-recovery","Disaster Recovery",{"slug":15,"name":16},"platform-as-a-service","Platform as a Service (PaaS)",{"slug":18,"name":19},"reserved-instances","Reserved Instances",{"slug":21,"name":22},"single-tenant-deployment","Single-Tenant Deployment",[24,28,32,35,38,41,45,48,51,54,57,60],{"slug":25,"category":5,"name":26,"updated_at":27},"air-gapped-deployment","Air-Gapped Deployment","2026-09-09T03:30:02+00:00",{"slug":29,"category":5,"name":30,"updated_at":31},"autoscaling","Autoscaling","2026-09-08T21:32:36+00:00",{"slug":33,"category":5,"name":34,"updated_at":31},"availability-zone","Availability Zone (AZ)",{"slug":36,"category":5,"name":37,"updated_at":31},"block-storage","Block Storage",{"slug":39,"category":5,"name":40,"updated_at":27},"bring-your-own-cloud","Bring Your Own Cloud (BYOC)",{"slug":42,"category":5,"name":43,"updated_at":44},"budget-alert","Budget Alert","2026-09-08T21:32:37+00:00",{"slug":46,"category":5,"name":47,"updated_at":27},"change-freeze","Change Freeze",{"slug":49,"category":5,"name":50,"updated_at":27},"control-plane","Control Plane and Data Plane",{"slug":52,"category":5,"name":53,"updated_at":44},"cost-allocation-tag","Cost Allocation Tag",{"slug":55,"category":5,"name":56,"updated_at":44},"cost-chargeback","Cost Chargeback",{"slug":58,"category":5,"name":59,"updated_at":44},"cost-observability","Cost Observability",{"slug":12,"category":5,"name":13,"updated_at":31}]