Multi-cloud adoption is increasing: 73% of enterprises say they are now using two public clouds, and 26% are using three or more, according to a 2021 VMware benchmark study.

Yet, IT infrastructure was already complex — thanks to numerous applications, workloads, and integrations. Managing and supporting systems across different clouds, as well as edge and on-premises data centers, only further complicates matters for IT departments. What’s more, in today’s highly competitive global business world, enterprises cannot afford business disruptions, downtime, and slow app performance.

But there’s good news: It’s possible to reduce complexity with improved integration from the data center to the cloud and edge locations using a stepping-stone approach.

The many benefits of multi-cloud

There are many reasons to have a multi-cloud strategy. Organisations are reaping the rewards of cost optimisations, avoiding vendor lock-in, cloud arbitrage, improved application performance, and more.

Another benefit is insulation from any problems a hyperscaler may have. Even the largest cloud providers like Microsoft, Google, and AWS suffer from outages. A multi-cloud architecture enables organisations to transfer workloads when a hyperscaler is having trouble.

Multi-cloud also gives IT departments the ability to place workloads as geographically close to business users as possible. For example, if you have US offices or employees in different states, you can use cloud providers in those regions to improve connectivity performance and adhere to data sovereignty issues.

These benefits translate to real-world business value. For example, the VMware Benchmark study found that among companies that have adopted multi-cloud:

  • 41% report reduced costs and IT hours spent on infrastructure
  • 35% have increased revenue from faster app delivery
  • 35% cite productivity savings across their distributed workforces

The challenges are real

All these benefits require overcoming some challenges, however. Among the top obstacles: Achieving visibility across compute, network, storage, and operating environments. In fact, 79% of enterprises said their cloud infrastructure is causing greater visibility gaps, according to a study by Axonius and Enterprise Strategy Group. In turn, these gaps can increase security and operational risks, including data breaches and/or the ability to achieve business objectives.

In addition, many IT departments are encountering challenges around joining legacy systems with cloud-native environments. These integration challenges occur at several levels. First, different hyperscalers require different skillsets or expertise. Each has different technologies, and all use different means of high-speed connection. The challenge for network architects: to develop an interconnecting mesh that provides a consistent backbone across all hyperscalers in their multi-cloud environment.

Another integration challenge involves developing a standard methodology for applications and workloads to communicate across clouds. In addition to using a standard protocol such as Representational State Transfer (REST), an organisation needs a consistent format and language that all applications use. In other words, architects need to develop a layer that ensures an Application Programming Interface (API) level of consistency between technologies such as AWS’s Lambda and Google Cloud Functions.

The list of risks associated with poor integration is a long one, including inconsistent infrastructure and siloed data, systems, and departments. Nonexistent integration can also cause costs to spiral, poor application performance for end users, and a drag on innovation efforts. All these slow down delivering on key objectives and creating business value.

Addressing multi-cloud issues

Many of the multi-cloud challenges that businesses are running into have to do with their approach, said Kit Colbert, vice president and CTO, Cloud Platform Business Unit, VMware.

“Essentially, they’re taking an ‘all or nothing approach’ and trying to go directly to a cloud-native nirvana or their future-state architecture in one fell swoop,” he said. “It is a monumental effort to rehost, refactor, and replatform all at the same time, so rather than going fast, this approach actually takes a lot longer and can fail.”

Instead, Colbert recommended a stepping-stone approach. Start by creating a holistic, multi-cloud vision for the entire organisation. This should act as a standard language for functions like wire protocols and APIs. As it evolves, the vision will become an organisational philosophy that considers, for example, operations, error handling, disaster recovery, compliance with regulations, and so on.

Next, take iterative steps toward multi-cloud nirvana. For example, start with a low-effort step like rehosting apps, then move to replatforming, and if needed, refactoring.

A single network architecture helps support this journey and brings consistency, efficiencies, and improved operations to your multi-cloud environment. Seek a horizontal approach, rather than a vertical, siloed cloud strategy, he notes, as this will allow for greater integration.

In addition, build or deploy a multi-cloud architecture that has:

  1. Flexibility. Your multi-cloud environment should increase overall business agility. That means creating an architecture that allows for each app to be deployed in the cloud that best meets its requirements.
  2. Portability. Your architecture should support the ability to move apps between environments — from one cloud to another — without having to refactor.
  3. Simplicity. Use a unified operating model to create a multi-cloud environment that reduces IT complexity and improves security.
  4. Optimisation. Overcome the cloud skills shortage by architecting a multi-cloud environment that leverages the expertise of existing IT teams and developers, while minimizing their learning curve.
  5. Automation. Build an architecture that supports your ability to scale up automation across all operations.

A foundation like VMware’s Unified Cloud Architecture allows organisations to take an iterative, layered approach to a streamlined multi-cloud environment. It offers the flexibility to build, run, and manage applications on any combination of clouds.

To learn more, visit