Solita Cloud Buyer’s Guide

2. Cloud service delivery models

There are five different cloud service delivery models, based on the procurement method and the level of responsibility. You can decide: how much you want to manage yourself and how much to buy as a service?

2. Cloud service delivery models

Cloud service delivery models

  1. Infrastructure as a Service (IaaS)

    Infrastructure as a Service (IaaS) refers to a service that would traditionally be a server solution you deploy in your own data center (on-premises), for instance storage and virtual machines.

    An IaaS service allows you to easily add capacity and the hardware is maintained by the cloud provider. The costs and features of different clouds are largely similar when it comes to IaaS capacity. Some differences do exist, but they may not be very relevant for choosing a cloud provider. The pricing is affected by the size of the server instances, the capacity type (on-demand or reserved), the operating time (start-stop), the architecture, the use of containers etc.

    Overall, with IaaS businesses avoid investing in expensive on-site resources up-front.

    Examples of IaaS are Amazon EC2, Azure Virtual Machines and Google Compute Engine.

  2. Container as a Service (CaaS)

    CaaS was not originally in the cloud service delivery model but has been added later. Although containers can be deployed using IaaS, with CaaS you don’t need to worry and manage container infrastructure yourself. In CaaS container infrastructure maintenance, patching and so on are managed by Cloud platform providers. CaaS is a subset of serverless*.

    Designed correctly, CaaS won’t end you up in a vendor-lock. Moving applications from cloud to cloud or from cloud to on-premises is technically possible.

    Examples of CaaS include: AWS Elastic Kubernetes Service (EKS), Azure Kubernetes Service (AKS) or Google Kubernetes Engine (GKE).

    • What does serverless mean? Serverless computing allows developers to build and run applications and services without thinking of virtual machines or physical servers. Despite the name “serverless”, there are still servers in the background but developers do not need to be aware of them. Example of serverless is e.g. Amazon RDS which is a database service provided without the need of server management or know-how. Second example of serverless is AWS Lambda. It is a function that developer can run to execute peace of code without the need of server management or know-how.
  3. Platform as a Service (PaaS)

    Platform as a Service (PaaS) refers to the virtual provision of computing resources over the cloud. This means it is primarily used by developers who are building software or applications.

    Platform-based application development is easier and faster, as the provider will handle all service updates and any functional complexities. For example, a platform will typically handle the adding of computing capacity according to operating loads. PaaS is a subset of serverless.

    PaaS services are crucial when vendor-lock is considered. PaaS components will by and large lock you in with the platform’s supplier. This may also be an advantage, as building your service with ready-made components is highly cost-effective compared to building a similar setup all by yourself.

    Google, AWS and Microsoft focus on continuously developing new cost-effective PaaS services. As a result, services related to databases, analytics, machine learning and the Internet of Things have improved significantly in the past few years and the competition between the three hyperscalers ensures that the development will continue in the times ahead.

  4. Function as a Service (FaaS)

    Function as a Service (FaaS) is a rather new cloud service delivery model. It provides a platform allowing customers to develop, to run and to manage application functionalities.

    FaaS is a subset of serverless: where serverless is focused on any service category (be it compute, storage, database, etc.), FaaS is focused on the event-driven computing wherein application code, or containers, only run in response to events or requests.

    Google Cloud Functions, Azure Functions and AWS Lambda are examples of FaaS.

    FaaS has some interesting benefits: it allows you to focus more on code, not infrastructure, to only pay when you use them, and to allow you to scale up or scale down automatically and rapidly.


  5. Software as a Service (SaaS)

    Software as a Service (SaaS) refers to an end-user application that is provided as a service. It is a model that makes software available to users over the internet, as these applications run directly through your web browser. They do not require any downloads or installations.

    The number of different SaaS services is constantly growing. These are the most common cloud services used by companies. Email (Office 365, Gmail), other software (Google G Suite, Salesforce, Showpad) and different financial administration services such as e-banking are good examples.

    The SaaS model offers a completely maintenance-free service for the customer, as the supplier will assume all responsibility. The service may be built in the supplier’s own data center or a public cloud by, for example, using AWS PaaS components. ITaaS (IT as a Service) is similar to the SaaS model, as companies use it to buy a tailored solution as a service.

Summary: Cloud service models and a cake

Now we have covered all five different cloud service delivery models. The following illustration summarizes the key differences between each delivery model.

Solita Cloud Buyer's Guide image: cloud service models as a cake