The Magazine

Agenda Digitale: The Five Pillars of the Cloud According to NIST – A Compass for Businesses and Public Administrations

Source:


By Lokesh Vij, Professor of Cloud Computing Infrastructure, Cloud Development, Cloud Computing Automation and Ops and Cloud Data Stacks at OPIT – Open Institute of Technology

NIST identifies five key characteristics of cloud computing: on-demand self-service, network access, resource pooling, elasticity, and metered service. These pillars explain the success of the global cloud market of 912 billion in 2025

In less than twenty years, the cloud has gone from a curiosity to an indispensable infrastructure. According to Precedence Research, the global market will reach 912 billion dollars in 2025 and will exceed 5.1 trillion in 2034. In Europe, the expected spending for 2025 will be almost 202 billion dollars. At the base of this success are five characteristics, identified by the NIST (National Institute of Standards and Technology): on-demand self-service, network access, shared resource pool, elasticity and measured service.

Understanding them means understanding why the cloud is the engine of digital transformation.

On-demand self-service: instant provisioning

The journey through the five pillars starts with the ability to put IT in the hands of users.

Without instant provisioning, the other benefits of the cloud remain potential. Users can turn resources on and off with a click or via API, without tickets or waiting. Provisioning a VM, database, or Kubernetes cluster takes seconds, not weeks, reducing time to market and encouraging continuous experimentation. A DevOps team that releases microservices multiple times a day or a fintech that tests dozens of credit-scoring models in parallel benefit from this immediacy. In OPIT labs, students create complete Kubernetes environments in two minutes, run load tests, and tear them down as soon as they’re done, paying only for the actual minutes.

Similarly, a biomedical research group can temporarily allocate hundreds of GPUs to train a deep-learning model and release them immediately afterwards, without tying up capital in hardware that will age rapidly. This flexibility allows the user to adapt resources to their needs in real time. There are no hard and fast constraints: you can activate a single machine and deactivate it when it is no longer needed, or start dozens of extra instances for a limited time and then release them. You only pay for what you actually use, without waste.

Wide network access: applications that follow the user everywhere

Once access to resources is made instantaneous, it is necessary to ensure that these resources are accessible from any location and device, maintaining a uniform user experience. The cloud lives on the network and guarantees ubiquity and independence from the device.

A web app based on HTTP/S can be used from a laptop, tablet or smartphone, without the user knowing where the containers are running. Geographic transparency allows for multi-channel strategies: you start a purchase on your phone and complete it on your desktop without interruptions. For the PA, this means providing digital identities everywhere, for the private sector, offering 24/7 customer service.

Broad access moves security from the physical perimeter to the digital identity and introduces zero-trust architecture, where every request is authenticated and authorized regardless of the user’s location.

All you need is a network connection to use the resources: from the office, from home or on the move, from computers and mobile devices. Access is independent of the platform used and occurs via standard web protocols and interfaces, ensuring interoperability.

Shared Resource Pools: The Economy of Scale of Multi-Tenancy

Ubiquitous access would be prohibitive without a sustainable economic model. This is where infrastructure sharing comes in.

The cloud provider’s infrastructure aggregates and shares computational resources among multiple users according to a multi-tenant model. The economies of scale of hyperscale data centers reduce costs and emissions, putting cutting-edge technologies within the reach of startups and SMBs.

Pooling centralizes patching, security, and capacity planning, freeing IT teams from repetitive tasks and reducing the company’s carbon footprint. Providers reinvest energy savings in next-generation hardware and immersion cooling research programs, amplifying the collective benefit.

Rapid Elasticity: Scaling at the Speed ​​of Business

Sharing resources is only effective if their allocation follows business demand in real time. With elasticity, the infrastructure expands or reduces resources in minutes following the load. The system behaves like a rubber band: if more power or more instances are needed to deal with a traffic spike, it automatically scales in real time; when demand drops, the additional resources are deactivated just as quickly.

This flexibility seems to offer unlimited resources. In practice, a company no longer has to buy excess servers to cover peaks in demand (which would remain unused during periods of low activity), but can obtain additional capacity from the cloud only when needed. The economic advantage is considerable: large initial investments are avoided and only the capacity actually used during peak periods is paid for.

In the OPIT cloud automation lab, students simulate a streaming platform that creates new Kubernetes pods as viewers increase and deletes them when the audience drops: a concrete example of balancing user experience and cost control. The effect is twofold: the user does not suffer slowdowns and the company avoids tying up capital in underutilized servers.

Metered Service: Transparency and Cost Governance

The dynamic scale generated by elasticity requires precise visibility into consumption and expenses : without measurement there is no governance. Metering makes every second of CPU, every gigabyte and every API call visible. Every consumption parameter is tracked and made available in transparent reports.

This data enables pay-per-use pricing , i.e. charges proportional to actual usage. For the customer, this translates into variable costs: you only pay for the resources actually consumed. Transparency helps you plan your budget: thanks to real-time data, it is easier to optimize expenses, for example by turning off unused resources. This eliminates unnecessary fixed costs, encouraging efficient use of resources.

The systemic value of the five pillars

When the five pillars work together, the effect is multiplier . Self-service and elasticity enable rapid response to workload changes, increasing or decreasing resources in real time, and fuel continuous experimentation; ubiquitous access and pooling provide global scalability; measurement ensures economic and environmental sustainability.

It is no surprise that the Italian market will grow from $12.4 billion in 2025 to $31.7 billion in 2030 with a CAGR of 20.6%. Manufacturers and retailers are migrating mission-critical loads to cloud-native platforms , gaining real-time data insights and reducing time to value .

From the laboratory to the business strategy

From theory to practice: the NIST pillars become a compass for the digital transformation of companies and Public Administration. In the classroom, we start with concrete exercises – such as the stress test of a video platform – to demonstrate the real impact of the five pillars on performance, costs and environmental KPIs.

The same approach can guide CIOs and innovators: if processes, governance and culture embody self-service, ubiquity, pooling, elasticity and measurement, the organization is ready to capture the full value of the cloud. Otherwise, it is necessary to recalibrate the strategy by investing in training, pilot projects and partnerships with providers. The NIST pillars thus confirm themselves not only as a classification model, but as the toolbox with which to build data-driven and sustainable enterprises.

Read the full article below (in Italian):

Related posts

Times of Malta: Malta-based OPIT launches innovative AI tool for students, academic staff
OPIT - Open Institute of Technology
OPIT - Open Institute of Technology
Sep 22, 2025 5 min read

Source:

4 min read


The launch was officially unveiled during an event held at Microsoft Italia in Milan, titled AI Agents and the Future of Higher Education.

A tech-focused higher education institution based and accredited in Malta has developed a new AI assistant designed to support both students and faculty.

In a statement, the Open Institute of Technology (OPIT), announced the launch of the OPIT AI Copilot.

With the Fall Term starting on September 15, OPIT said it has already launched beta testing with faculty champions and is currently piloting full-course integrations.

Students who will be part of the pilot-phase will be able to prompt the entire OPIT – Open Institute of Technology knowledge base, personalized to their own progress.

The platform was developed entirely in-house to fully personalize the experience for the students, and also make it a real-life playground for in-class projects. It is among the first custom-built AI agents to be deployed by an accredited European higher education institution.

The launch was officially unveiled during an event held at Microsoft Italia in Milan, titled AI Agents and the Future of Higher Education

The gathering brought together academics and technology leaders from prominent European Institutions, such as Instituto de Empresa (IE University), OPIT itself and the Royal College of Arts, to explore how artificial intelligence is reshaping the university experience.

The OPIT AI Copilot has been trained on the institute’s complete academic archive, a collection created over the past three years that includes 131 courses, more than 3,500 hours of recorded lectures, 7,500 study resources, 320 certified assessments, and thousands of exercises and original learning documents.

Unlike generic AI tools, the Copilot is deeply integrated with OPIT’s learning management system, allowing it to track each student’s progress and provide tailored support.

This integration means the assistant can reference relevant sources within the learning environment, adapt to the student’s stage of study, and ensure that unreleased course content remains inaccessible.

A mobile app is also scheduled for release this autumn, that will allow students to download exercise and access other tools.

During examinations, the Copilot automatically switches to what the institute calls an “anti-cheating mode”, restricting itself to general research support rather than providing direct answers.

For OPIT’s international community of 500 students from nearly 100 countries, many of whom balance studies with full-time work, the ability to access personalised assistance at any time of day is a key advantage.

“Eighty-five per cent of students are already using large language models in some way to study,” said OPIT founder and director Riccardo Ocleppo. “We wanted to go further by creating a solution tailored to our own community, reflecting the real experiences of remote learners and working professionals.”

Tool aims to cut correction time by 30%

The Copilot will also reduce administrative burdens for faculty. It can help grade assignments, generate new educational materials, and create rubrics that allow teachers to cut correction time by as much as 30 per cent.

According to OPIT, this will free up staff to dedicate more time to teaching and direct student engagement.

At the Milan event, Rector Francesco Profumo underlined the broader implications of AI in higher education. “We are in the midst of a deep transformation, where AI is no longer just a tool: it is an environment that radically changes how we learn, teach, and create,” he said.

“But it is not a shortcut. It is a cultural, ethical, and pedagogical challenge, and to meet it we must have the courage to rethink traditional models and build bridges between human and artificial intelligence.”

OPIT was joined on stage by representatives from other leading institutions, including Danielle Barrios O’Neill of the Royal College of Art, who spoke about the role of AI in art and creativity, and Francisco Machin of IE University, who discussed applications in business and management education.

OPIT student Asya Mantovani, also employed at a leading technology and consulting firm in Italy,  gave a first-hand account of balancing professional life with online study.

The assistant has been in development for the past eight months, involving a team of OPIT professors, researchers, and engineers.

Ocleppo stressed that OPIT intends to make its AI innovations available beyond its own institution. “We want to put technology at the service of higher education,” he said.

“Our goal is to develop solutions not only for our own students, but also to share with global institutions eager to innovate the learning experience in a future that is approaching very quickly.”

Read the full article below:

Read the article
E-book: AI Agents in Education
OPIT - Open Institute of Technology
OPIT - Open Institute of Technology
Sep 15, 2025 3 min read

From personalization to productivity: AI at the heart of the educational experience.

Click this link to read and download the e-book.

At its core, teaching is a simple endeavour. The experienced and learned pass on their knowledge and wisdom to new generations. Nothing has changed in that regard. What has changed is how new technologies emerge to facilitate that passing on of knowledge. The printing press, computers, the internet – all have transformed how educators teach and how students learn.

Artificial intelligence (AI) is the next game-changer in the educational space.

Specifically, AI agents have emerged as tools that utilize all of AI’s core strengths, such as data gathering and analysis, pattern identification, and information condensing. Those strengths have been refined, first into simple chatbots capable of providing answers, and now into agents capable of adapting how they learn and adjusting to the environment in which they’re placed. This adaptability, in particular, makes AI agents vital in the educational realm.

The reasons why are simple. AI agents can collect, analyse, and condense massive amounts of educational material across multiple subject areas. More importantly, they can deliver that information to students while observing how the students engage with the material presented. Those observations open the door for tweaks. An AI agent learns alongside their student. Only, the agent’s learning focuses on how it can adapt its delivery to account for a student’s strengths, weaknesses, interests, and existing knowledge.

Think of an AI agent like having a tutor – one who eschews set lesson plans in favour of an adaptive approach designed and tweaked constantly for each specific student.

In this eBook, the Open Institute of Technology (OPIT) will take you on a journey through the world of AI agents as they pertain to education. You will learn what these agents are, how they work, and what they’re capable of achieving in the educational sector. We also explore best practices and key approaches, focusing on how educators can use AI agents to the benefit of their students. Finally, we will discuss other AI tools that both complement and enhance an AI agent’s capabilities, ensuring you deliver the best possible educational experience to your students.

Read the article