Multi-cloud is using infrastructure or platform services from two or more public cloud providers, such as Amazon Web Services (AWS), Microsoft Azure, Google Cloud or Oracle Cloud, within the same organization. Sometimes it is a deliberate strategy to use the best service for each job or to avoid depending on one provider. Often it simply happens: different teams pick different clouds, an acquisition brings another provider, or a key application only runs on one platform. Either way, the organization has to manage identity, networking, security, cost and skills across more than one provider.
At a glance
- Multi-cloud means two or more public cloud providers in use; it is different from hybrid cloud, though one environment can be both.
- Common reasons are best-fit services, reduced dependence on one provider, regional coverage, acquisitions and customer or regulatory requirements.
- Running one application across clouds for resilience is possible but uncommon and harder than running different applications on different clouds.
- The main costs are added complexity, data transfer charges between providers and duplicated tools and skills.
- Consistent identity, networking, security policies and cost reporting across providers make it manageable.
What problem it solves
Relying on a single cloud provider concentrates risk. Your pricing, service availability, regional footprint and roadmap all depend on one company, and moving away later becomes harder the more of its proprietary services you use. Some workloads also run better on a particular provider, because of a database, analytics or AI service, a regional presence, or licensing advantages for software you already own.
Multi-cloud lets you choose per workload and keeps alternatives open. For companies that grew through acquisition or let teams choose freely, the problem is the reverse: they already have several clouds, and need a way to govern them consistently instead of as separate islands.
How it works
Placement. Most multi-cloud estates place each application on the provider that suits it, rather than splitting one application across providers. Spreading a single application across clouds, or keeping a standby copy on a second provider, is possible but requires portable designs, often based on containers and Kubernetes, data replication and regular testing.
Identity and governance. A central identity provider with single sign-on across every cloud’s console, plus consistent account structures, tagging and policies, prevents each cloud from becoming its own silo. Security posture tools such as cloud security posture management (CSPM) check configurations across providers.
Networking. Clouds need to reach each other, your offices and your data centers. Options include VPNs over the internet, private connections into a colocation hub or carrier network, cloud-interconnect and network-as-a-service platforms, and SD-WAN extended into each cloud. Data moving out of a cloud usually carries transfer charges, though rates and discounts vary by provider and connection type, so architecture choices directly affect cost. Our blog on how SD-WAN supports multi-cloud environments covers one approach.
Cost management. Each provider bills differently. FinOps practices and tools bring spend from every cloud into one view and allocate it to teams, while commitment discounts need planning per provider. More on our Multi-Cloud solutions page.
When it matters for buyers
- After a merger or acquisition. Inherited cloud accounts on a different provider force a decision: consolidate, or govern both.
- When negotiating cloud commitments. A credible second provider can improve terms, but splitting spend can also reduce volume discounts.
- When expanding into new regions. Regional coverage and data-residency requirements may favor a different provider in some countries.
- When a customer or regulator asks about concentration risk. Some industries are expected to show they can cope with a cloud provider failure.
- When cloud costs are hard to explain. Inter-cloud data transfer and duplicated tools often show up here.
Questions to ask vendors
- What does data transfer out of your cloud cost, to the internet and to other clouds, and are private connections cheaper?
- Which of your services are proprietary, and what would it take to move off them?
- Do your security, monitoring and cost tools support other clouds, and at what extra cost?
- How do commitment discounts work if our spend is split across providers?
- Which interconnect locations and partners do you support for private connectivity?
- What support do you offer for workloads that span your cloud and others?
How it differs from hybrid cloud
Multi-cloud and hybrid cloud describe different things. Multi-cloud is about how many public cloud providers you use. Hybrid cloud is about combining private infrastructure, such as your own data center, colocation or a private cloud, with public cloud so they operate together, with shared networking, identity and management. A company can be multi-cloud without being hybrid (all workloads on AWS and Azure), hybrid without being multi-cloud (a data center plus one public cloud), or both at once.
Problems this helps with
- Different teams running on different clouds with no common controls
- Being dependent on a single cloud provider's pricing and roadmap
- Data transfer charges between clouds we didn't plan for
- Security and identity settings that don't match across providers
