Multi-cloud is an IT strategy in which an organization uses more than one cloud service from more than one service provider. Examples of multi-cloud configurations include one public and one private cloud, two public and one private cloud, and three or more public clouds.
Multi-cloud clusters — a feature available in MongoDB Atlas, a global cloud database service — takes the concept a step further by enabling a single application to use multiple clouds. With multi-cloud clusters, data is distributed across different public clouds (Amazon Web Services (AWS), Google Cloud Platform (GCP), Microsoft Azure), enabling deployment of a single database across multiple providers simultaneously.
No one cloud delivers all things to all people. Cloud vendors offer different blends of service-level agreements, availability, security architectures, pricing, tools, and resources for various needs. An increasing number of enterprises will cherry-pick among the various cloud vendors to find the best combination of features for each specific workload. Multi-cloud is increasingly prevalent, with 86% of organizations around the world describing their strategy as multi-cloud. For enterprises with 1,000 or more employees, the figure is nearly the same, at 84%. In addition, organizations report using an average of five clouds.
Application architectures across multiple clouds don’t have to be linked to one another, but as they become more complex, the need arises for cloud-agnostic centralized management. This requires considerable in-house expertise in cloud infrastructure, administrative automation, and management tools.
Enterprises of all sizes adopt a multi-cloud strategy to achieve a variety of goals:
IT departments often find themselves with a multi-cloud not of their own making, owing to business units adopting technology on their own. When brought to IT’s attention, such “rogue” clouds are generally brought under centralized oversight.
It’s smarter to choose a cloud with the best combination of features for each application than to tinker with the workloads themselves: a “lowest common denominator” application will likely shortchange high performance to achieve cloud portability. For example, the best option for a given company may be to use one cloud provider for infrastructure, another for development, and a third for performance tools.
An almost universal driver for multi-cloud adoption is the desire to avoid being confined to any single provider’s infrastructure, pricing model, and specialized services.
No cloud provider is completely immune from outages. Taking a multi-cloud approach, especially one utilizing backup clouds, provides protection from the risk of having a business-critical application become unavailable.
Data governance requirements, such as the EU's GDPR, sometimes stipulate that customer data must be held in particular locations. This can require a multi-cloud strategy if your primary cloud does not support all the regions where your customers reside.
Latency can be minimized by choosing a cloud provider with a presence geographically close to customers. For organizations serving users in multiple regions, the optimal solution may involve multiple cloud providers.
Distributing your workloads and traffic among a number of clouds reduces the risk of distributed denial of service (DDoS) threats.
Multi-cloud clusters solve one of the fundamental issues facing organizations today: data portability. While code is increasingly cloud agnostic and easy to move to the best environment, data is not.
By solving the data portability issue, multi-cloud clusters grant organizations additional benefits over other multi-cloud setups, such as:
There’s no need to manually pipe data into different cloud ecosystems to take advantage of unique tools and services, such as AWS Lambda functions, Google Cloud AI/ML offerings, or Azure IDE tools.
Offer end-customers a choice of what cloud ecosystem to use and store data in, easily satisfying their preference and compliance requirements. Also, distribute data in more regions and reach users in areas where only one cloud provider is present.
Move your data from one cloud to another, with no downtime.
With Atlas multi-cloud clusters, an application can be configured to survive a cloud provider outage or downtime due to regional capacity constraints.
Learn more about multi-cloud clusters on MongoDB Atlas →
Multi-cloud contains cloud deployments sourced from different vendors. In contrast, a hybrid cloud indicates the presence of both deployment types — public and private — essentially functioning as a single unit, with orchestration tools used to deploy and manage workloads between the two components. In recent years, the line between the two cloud types has been blurring. An increasingly common model is the hybrid multi-cloud, utilizing a private cloud plus several public clouds from different service providers.