Cloud Computing: Models, Issues, and Advantages

Cloud computing refers to computations performed not on the user's local machine, but within a cluster of remote servers, i.e., "in the cloud". 

Large clouds can be geographically distributed, meaning groups of servers are located in different data centers connected to each other. 

This is the essence of cloud computing—computation is based on the collective (as a unified whole) use of resources from separate physical machines. 

As a result, the payment model is based on actual usage of resources in real-world tasks. You pay as much as you use—almost a linear dependency.

The standard set of cloud services includes:

• computing power,
• data storage,
• administration and development tools (optional),
• additional services and applications (optional).

Cloud capabilities and terms of use depend on the specific provider that owns the physical servers.

Cloud in the Network Infrastructure Context

When the internet was just beginning, the concept of the cloud did not exist. In the 2000s, the average company could have its own small servers (a mini data center), maintained by a couple of specialists. This was a common approach. 

However, as global network infrastructure became more complex, centralized hubs emerged—companies owning a large number of physical servers in one location and renting them out. 

This network’s “evolution” into such an architecture is essentially the result of the division of labor, especially in the context of globalization. Everyone focuses on their own expertise: some manage networks and servers, others develop services and applications.

From an infrastructure standpoint, however, the cloud is more than just server rental. It’s a broader concept. 

A cloud is more flexible. All physical servers of a provider are combined into a single entity, fragments of which are provided to clients. Hence, higher flexibility and scalability, instantaneously on demand.

Cloud Computing Models

The cloud clearly separates physical and software levels. This high level of abstraction is the hallmark of cloud computing. The entire range of clouds, cloud services, and applications built on them can be divided into several software layers:

• IaaS (Infrastructure as a Service): This layer provides only computing resources. This is pure cloud. Examples include Amazon Web Services and Elastic Compute Cloud. However, since the 2010s, IaaS has lost popularity compared to PaaS and SaaS models.

• PaaS (Platform as a Service): This layer offers not just the cloud, but a full platform with preinstalled development and administration tools. It provides a complete stack for specific application tasks. This can include operating systems, databases, containers, libraries, frameworks, microservices, etc. Examples include Google Cloud Platform and Red Hat OpenShift, the latter offering containerization tools like Docker and Kubernetes.

• SaaS (Software as a Service): This layer includes client-facing cloud services with ready-to-use functionality that is accessible "out of the box." Users typically pay a monthly or annual subscription fee, although free versions with limitations are often available. SaaS is by far the most widespread model. Examples: Google Drive, Salesforce, Dropbox.

From the user’s perspective, cloud computing thus represents serverless computing—a model where infrastructure management is shifted to the provider, allowing developers to focus on business logic. This approach enables automatic scaling based on request volume, with application logic executed only when needed—users pay for actual computations, not idle capacity.

This fourth model, in addition to the three above, is called FaaS (Function as a Service). FaaS allows developers to execute code (functions) in response to specific events—triggers. Charges apply only when the code runs.

Cloud Types

The cloud model is not the same as the cloud type. The model relates to how the infrastructure is organized and how the provider interacts with clients (technically and economically). The type describes how the cloud integrates into the broader internet infrastructure, i.e., how strongly or weakly the provider's physical servers are differentiated within the global network.

Here are the main types of cloud:

Public Cloud

The most common type. Most clouds—especially SaaS—are public, open to everyone and not isolated from the internet. Any user can access remote servers via an interface, typically a browser or client app; for example, Google Drive via a web page, Dropbox via a desktop app.

These clouds often follow a subscription model offering extended features. If intended for personal users, a limited free version is usually available. For enterprise clients, everything must be prepaid. 

Public clouds typically serve many clients, often individuals. Services like AWS, Google Cloud, IBM Cloud, Microsoft Azure, and Oracle Cloud have millions of users.

The provider is responsible for the data centers. Users expect strong security, achieved through internal corporate controls and encryption. Still, data breaches occur. With public clouds, there’s no absolute guarantee of security.

Hence, corporate clients often prefer other cloud types.

Private Cloud

A cloud environment where computing resources are entirely allocated (isolated) for a single client: individual or corporate. It is as flexible and scalable as a public cloud but without the drawback of public access. It's harder to hack and is controlled directly by the client. It may be hosted locally or by a provider, but server capacities may be segregated per client.

The main reason for using a private cloud is the protection of confidential documents, intellectual property, financial, and personal data.

However, private clouds require expertise—specialists for deployment and management. For young projects, this may be costly.

Hybrid Cloud

A hybrid cloud combines public and private resources to achieve greater flexibility and security. Such clouds orchestrate workloads depending on the task: sensitive data remains in the private segment (more expensive to maintain), and business logic is in the public segment (where high control is not needed).

This approach allows organizations to meet goals more efficiently than relying on a single type. With a custom private cloud architecture, a company can move workloads to the public cloud when needed, optimizing resources.

Multicloud

A multicloud is an abstraction over multiple clouds from different providers. These are different clouds, united in one ecosystem to perform shared tasks. It can combine SaaS, PaaS, and IaaS models.

For example, email services may come from one provider, a code repository from another, and media storage from a third. Multiclouds can also be hybrid, combining public and private environments.

Companies benefit from multiclouds by avoiding dependence on a single vendor, like diversification. They can switch to more innovative providers when needed. On the flip side, managing many tools complicates the environment. It's important to strike a balance—too many dependencies can backfire.

Advantages of Cloud Computing

Compared to on-premise server infrastructure or renting dedicated resources, cloud computing offers clear advantages:

• Cost Reduction

Clouds maintained by global network participants relieve developers (users) from most expenses. No need to buy, install, configure, or manage servers. No need for deep server management expertise. Development companies use ready-made clouds, keeping only essential specialists focused on the product and business logic. This significantly reduces both financial and time costs.

• Flexibility and Speed 

Provider-supplied clouds typically include server applications and services that assist in development and administration. Access is instant. Options are immediately available—no waiting on traditional IT teams or vendors. Some providers even allow developers to customize their cloud segments.

• Efficient Scalability 

Clouds can be resized depending on business logic needs. Instead of overprovisioning, the application scales up or down based on load, in real time, within the user request pipeline.

Cloud Computing Challenges

• Security and Confidentiality 

Most cloud issues relate to privacy. Public cloud providers can access stored data: code, media, databases. Large providers focus on security and usually have no interest in user data.

However, smaller, less-known providers may sell or misuse data. Some disclose this in privacy policies that users agree to. Some contracts omit data ownership rights entirely—legal action may be ineffective. Thus, reputation and brand size matter. Be cautious with unknown vendors.

• Insecure Interfaces 

The client-provider connection is a critical security point. Proper credential control and encryption protocols are essential. Providers must build safe, user-friendly API architectures.

Many users lack understanding of risks when connecting to remote servers. Security literacy is vital, especially as cloud computing becomes routine. What's the point of encrypting everything if you store passwords in a plain .txt file?

• Outages and Data Loss 

Providers can experience issues beyond the client’s control. To clients, servers are a "black box." Failures can cause anything from data loss to complete computation halts.

• Dependency and Support Guarantees

Except for private clouds, users don’t own the infrastructure—the provider does. Outsourced infrastructure creates dependency. Small or mid-size providers may not guarantee longevity. Major providers (Google, Amazon) are more reliable, but users must comply with their policies and limitations. In this sense, physical control over equipment (as with private clouds) is preferable, albeit costlier and more demanding.

Conclusion

The key feature of cloud computing—the separation of physical and logical layers—has become a driving force in the evolution of modern network infrastructure. 

The global network is moving toward a state where devices around the world connect to localized hubs: data centers or clouds. 

This pervasive network is now known as the Internet of Things (IoT), a concept where every object is connected to millions of others through a global shared network. It's a world where even your vacuum and coffee machine have digital reflections (avatars) on the internet. The physical world is mirrored in virtual reality.

In the end, the global internet is turning into a distributed network of localized clouds, connected and communicating with one another. 

Yes, the internet has always been decentralized. But the cloud era brings greater flexibility, adaptability, and bandwidth. Such a network is far more responsive to change than traditional architecture. And this global cloud trend is only gaining momentum.