While it is no longer news, cloud computing continues to grow. In Italy, the past year saw a 24% increase in spending (PoliMI Observatory), primarily in the Public & Hybrid Cloud space and, more specifically, in infrastructure services (IaaS), which have now reached the historically dominant software services (SaaS) sector.
In many companies, the cloud has changed everything: the provisioning of IT resources, how applications are developed and managed, the technologies companies can rely on, security paradigms, and resource access from anywhere with any device. As if that weren't enough, it has also transformed how businesses operate and manage internal and external relationships.
One of the disruptive aspects of the cloud, particularly in its public form, has always been the pricing model, which, besides being profoundly different from the past and much more flexible, also brings some specific risks. This is why cloud cost management has become an essential discipline for any modern enterprise.
Despite some exceptions, such as fixed pricing and discount policies based on certain customer commitments, the pricing logic for public cloud services is pay-per-use. Essentially, companies pay for what they use, with a periodic billing cycle.
Pay-per-use, closely tied to the as-a-service approach that characterizes the entire cloud ecosystem, has long been considered a significant benefit of the new IT paradigm. Since businesses only pay for what they consume, even those with limited resources can access cutting-edge technology, enhance their competitiveness, and contribute to systemic innovation, all while paying only for what they need.
In practice, however, tracking and optimizing cloud resource costs is a real challenge for two key reasons:
In recent years, a phenomenon known as cloud repatriation has emerged. Some companies that were initially very active in transitioning to the public cloud are now reshaping their IT architecture and increasingly leveraging a hybrid or private approach. There are several reasons for this, ranging from performance to compliance. However, better cost management undoubtedly plays a crucial role —not just in terms of reducing expenses but in optimizing technological investments. In other words, companies are trying to balance costs with the value obtained and reduce the risk of unexpected expenses associated with the intensive use of public cloud resources.
Without in-depth knowledge of pricing structures, which some company divisions lack, the risk of overprovisioning cloud resources is very concrete, especially given the self-service nature of the services used. This can lead to underutilization of multiple resources, which, despite not being used to their full capacity, continue to generate costs for the company, impacting overall efficiency.
Further complicating the situation is the lack of visibility into spending, a problem that stems not only from the inherent complexity of tariffs and consumption patterns (although every provider offers tools for analyzing and managing consumption) but also from the increasing adoption of multicloud strategies by enterprises. These organizations use services and resources from various providers and integrate them into a single IT ecosystem, but this fragmentation makes it even harder to understand and, most importantly, reliably predict costs.
Cloud cost management is thus a discipline that aims to monitor, manage, and optimize the costs associated with the use of multiple cloud services supporting enterprise data and applications. As previously mentioned, the goal is not merely to reduce expenses but to ensure that resources are used efficiently and in line with the company's business objectives.
To effectively manage cloud costs, organizations must adopt a holistic approach that includes both tactical and strategic aspects of financial optimization, given their complementary nature. The former, in particular, focuses on short-term actions and rapid resolution of inefficiencies to achieve immediate results. Below are some common techniques.
Organizations often have allocated computing or storage instances that are not actively being used, representing a direct waste of resources and unnecessary cost.
A flexible paradigm like the cloud allows resources to be automatically increased or decreased based on actual usage needs.
A common feature of hyperscaler offerings is reserved instances. This involves purchasing cloud resources with a commitment to using a certain amount over a predefined period in exchange for significantly lower prices compared to on-demand instances.
Not all instances need to be available 365 days a year. Shutting them down when not needed is a basic rule that not everyone follows.
Solutions and platforms specifically dedicated to cloud cost management are available on the market. They provide visibility into complex ecosystems (including hybrid and multicloud) and reporting, but also AI-based suggestions for optimizing overall spending based on the required performance.
An interesting technique aimed at maximizing visibility, and thus cost optimization, is the assignment of metadata to the cloud resources and services used. This makes it easier to monitor costs with breakdowns, for example, by department, cost center, or project.
The strategic approach to cloud cost optimization focuses on long-term management aimed at ensuring a balance between the efficient use of resources, scalability, and cost containment. In any case, an in-depth analysis of usage trends and future organizational requirements is necessary, as well as the use of tools, methodologies, and frameworks that proactively and sustainably optimize costs.
In the cloud era, capacity management is a dynamic and predictive process. The key is the accurate analysis of historical usage data combined with business growth forecasts, to create models that accurately estimate future cloud resource needs. In addition to the obvious benefits of resource allocation, this allows companies to negotiate more favorable long-term contracts with cloud providers.
Just as DevOps revolutionized the way companies develop and manage complex software projects, FinOps is a framework that redefines the entire concept of financial cloud management. It's no coincidence that FinOps’ key principle is the integration of financial, technical, and operational teams to make informed decisions aligned with the business and the overall company goals. FinOps, in other words, is an antidote to the fragmentation of cloud resource purchasing.
Adopting a multicloud strategy offers advantages that go beyond merely avoiding vendor lock-in. This approach allows organizations to diversify risk by distributing data, files, and workloads across multiple providers, leveraging the most competitive offers, and optimizing costs by aligning each application or service with the cloud provider best suited to its needs.
The Cloud Center of Excellence (CCoE) is a cross-functional team representing multiple business areas (IT, finance, security, operations...), tasked with defining best practices and cloud governance rules. For instance, the CCoE sets guidelines for resource use, determines who has the authority to acquire them, and ensures their security. The Center of Excellence also defines best practices for resource optimization and establishes which tools should be used for monitoring and reporting.