Why multi-cloud rationalization matters

Multi-cloud — where organizations use a combination of different cloud platforms and services — is surprisingly common. According to the Flexera State of the Cloud Report 2022, 89% businesses operate multi-cloud. From these, 80% use hybrid clouds, 7% use only multiple public clouds and 2% use only multiple private clouds.

Although multi-cloud can offer significant benefits to organizations — the State of DevOps report 2021 suggests that firms using hybrid or multi-cloud are 1.6 times more likely to exceed their organizational performance targets — it requires rationalization to be truly effective and impactful. In other words, it needs to be properly aligned to an organization’s business strategy and goals. Indeed, without rationalization multi-cloud can introduce risks to a business. These include hampering operational effectiveness, increasing costs, risking non-compliance and even making it vulnerable to security threats. These issues often arise when loose governance leads to accidental multi-clouds or where circumstances (like, say, a merger) lead to multi-cloud compromises. 

However, even the most intentional and well-planned multi-cloud projects can go astray without adhering to the guiding principles and governance. Without continuous realignment with business strategy and outcomes, technology guiding principles, and governance, intentional multi-clouds can accumulate accidental and optional complexity which increases technical debt and reduces business impact.

This blog post explains how multi-cloud rationalization can be done. It will help you successfully navigate some of the risks of multi-cloud and ensure that your organization is well-set to leverage its many advantages.

Why do organizations fail to maintain intentional multi-cloud systems?

Before we explore how to properly rationalize your multi-cloud approach, it’s worth reflecting on some of the ways that organizations arrive at multi-clouds in apparently unintentional or accidental ways.

Accidental multi-clouds

What we call accidental multi-clouds typically emerge when organizations lack a cohesive cloud adoption strategy or one that has not been effectively socialized and adhered to. One of the most obvious symptoms is cloud sprawl and an unnecessarily complex technology estate. In turn, this increases technical debt and escalates business risks and costs. This lack of cohesion is usually caused by a lack of effective guidance or governance, with teams operating independently and autonomously, often dictated by delivery pressures of localized necessities, without a unifying strategy. 

Consequential multi-clouds

Another issue can arise when cloud systems come about through the consequences of a merger or acquisition. These are what we call consequential multi-clouds. 

In these scenarios, two different firms will come together with their own respective technology built on different public and private clouds. When faced with time pressures (not uncommon in an M&A process) businesses will often try to integrate their systems under a common abstraction layer. Underneath such layers, these systems largely operate independently yet still have huge overlaps in terms of their capabilities. This leads to different customer segments being served by different systems. Again, without the required levels of guidance and governance to align a cloud system with organizational structures and goals post M&A, this creates increased complexity and increased costs.

The dangers of a lack of rationalization

The most obvious challenge of multi-cloud is technological divergence. This places a burden on engineering teams — when capabilities that are available on one platform aren’t available on another, it will be down to them to install the necessary services and appropriately configure, maintain and operate the system. 

On even single cloud platforms, developers have to choose between multiple similar capabilities —  containerization on AWS can be done with either EKS, ECS or Fargate. This gets even more complicated when you’re using multi-cloud, making governance absolutely essential.

Integrating services on different cloud platforms also increases complexity and overheads, making it harder to ensure uniformity in provisioning and delivery pipelines, security and compliance policy rollouts and monitoring and observability. In turn, this can make it harder to fulfill certain cross-functional requirements like performance, security and reliability.

Of course, technical complexity has further downstream consequences — it requires talent that is comfortable working across different platforms and tools, and has the knowledge and skills to help  optimize cloud costs, which potentially is more complicated in a multi-cloud scenario. Overall, this means greater organizational complexity. And while it’s true that multi-cloud rationalization will require organizational shifts and new structures, what’s different when you rationalize your approach is that you’re in control — you can be deliberate and intentional. Complexity is something you manage, not something that just happens to you.

An intentional and rationalized approach to multi-cloud

Multi-cloud without governance, oversight and strategic intention can be extremely troublesome for organizations. Not only will it prevent you from leveraging the substantial benefits that an effective multi-cloud approach can offer, it could do further damage — increasing your costs while slowing down your ability to innovate and deliver for customers quickly.

There are a few key actions that must be followed if you are to ensure a properly intentional and rationalized multi-cloud.

Risk management: Multi-cloud can increase risks for organizations. These need to be identified and steps should be taken to proactively mitigate them. For example, reconciling inconsistent identity and entitlements across multiple clouds that can create security issues. 

Simplicity: Adding multiple cloud technologies from different vendors will inevitably contribute to greater complexity. This is why it’s essential to follow architectural principles and designs that will reduce complexity. For example, although automatic failover across multiple clouds can, in theory, reduce the risk of outages, it can be very complicated to actually implement. Consequently, it should be avoided in favor of disaster recovery approaches of individual providers.

Portability commensurate with risk: Application portability should always be proportional to the risk you’re trying to mitigate. This is because portability can be costly. For example, banks mitigate concentration risk with portability of only critical business capabilities.

Fulfill cross functional requirements (CFRs) across cloud platforms: CFRs such as performance, security, cost and resilience, help achieve business outcomes alongside functional requirements. Adopting multi-cloud should never stop you from meeting those CFRs.

Promote a unified operating model for multi-cloud governance: Disparate governance models using diverse tools and processes increase costs, complexity and the risks of multi-cloud governance. An objective, federated and responsive governance model is essential to continuously rationalize a multi-cloud estate.

How to rationalize multi-cloud strategically

We’ve seen why rationalizing multi-clouds can be challenging and some of the risks this can pose and explored key actions and principles that should form the foundations of multi-cloud rationalization. However, what actually needs to be done?

Multi-cloud assessment

Integral to multi-cloud rationalization is to perform a regular assessment of your cloud estate. In particular, attention should be paid to how it’s delivering value to the business. It’s also important to be candid about costs and risks.

Technical approaches to rationalization

It’s essential to build a catalog of technical approaches, technologies and sensible defaults for multi-cloud rationalization. These should, however, be regularly reviewed; those involved must be cognizant of when something should be applied and the potential consequences. From a governance perspective, it can be useful to create an internal Technology Radar to guide technology decision making. Last but not least, an evolutionary engineering platform used to deploy and operate technology consistently and uniformly minimizes cognitive overload. This enables teams to deliver value with both speed and stability. 

Organizational considerations 

Multi-cloud requires a number of organizational considerations, such as training and talent development and organizational structure. Ideally, a technical organization should always have business value aligned teams that are supported by platform teams.

The impact on your operating model

A multi-cloud estate requires a more intricate operating model to ensure that it is not treated as multiple individual clouds. Multi-cloud rationalization means that this operating model needs to be adaptable and flexible.

Commercial impact

Commercial agreements, like committed usage discounts, with different cloud vendors will regulate multi-cloud rationalization. Rationalization initiatives will have to determine break even points for these contracts to ensure that rationalization remains cost effective for the business.

Multi-cloud can offer substantial benefits to businesses — but it needs to be managed intentionally with rationalization to ensure alignment with business objectives.  The investment required here has substantial dividends for both business and technology. Taking the necessary steps to rationalize your multi-cloud approach — whatever your context — will put you in the best possible position to gain maximum impact from it.