Function as a Service (FaaS)

In recent years, a number of IT solutions have been developed in the field of cloud computing that offer users advantages primarily through location-independent availability and flexible use. The probably best-known field, Software as a Service or SaaS for short, covers all types of software that can be used via an Internet connection without having to own a local version.

Function as a Service or FaaS describes application packages that are developed, managed, and executed by developers. At the same time, their own infrastructure isn’t required for their management, as it’s provided by the service provider. The development of the functions takes place in isolation in so-called containers. Data processing or mobile apps are common examples of functions used with FaaS. In this article, we explain which areas Function as a Service covers, how developers can use the service, and what the advantages and disadvantages of the cloud computing solution are.

Function as a Service is a cloud computing service with which applications can be developed and put into operation. The special feature is that the server on which the application runs and the infrastructure of the service are the responsibility of the provider, meaning that associated tasks like maintenance or regular updates are not necessary for the developers. This allows them to focus entirely on programming code for the required application packages. Here, the term “function” refers to a software component that is often executed for only seconds. Applications consist of one or several functions, and it’s also possible to assemble microservice components that consist of a large number of functions.

Although many of the FaaS solutions are cloud-based, some of the FaaS architectures are also deployed locally and are installed and hosted in purpose-built data centers. The provider charges for the use of the FaaS solution based on the computing power required and the data storage used.

Function as a Service provides developers with a method for implementing serverless computing. For this purpose, a business logic is written within a so-called container, and management is carried out entirely via a readily provided platform. This platform is usually used in the cloud, but the model is also already being used with local and hybrid deployments.

What’s special: FaaS follows a reactive system that is controlled by events. The events are triggered by specific occasions, such as the click on a button, which results in the function being called up. Once the call has been made, the service waits until a new event is generated. The basis for how FaaS works is that the response to the occasion is started and executed straight away. For example, the upload of an image file can trigger a reaction that results in the file being automatically converted to a specific format.

To execute the responses of the Function as a Service logic, an infrastructure is available at any time, so that permanent server processes aren’t needed in the background. This in turn has a positive effect on scaling options: The cloud provider ensures the availability of the defined functions and provides resource allocation to the second. As a result, when few requests arrive, the application is scaled down. With FaaS, you’re charged only for the resources used – standby times are therefore not incurred.

In practice, two types of applications are particularly well suited for FaaS implementation: infrequently executed workloads and high-volume transactions.

The fundamental difference between FaaS and SaaS lies in how the two cloud services are applied. While FaaS operates in the microcosm of the application landscape and is limited to simple functions, Software as a Service provides users with extensive applications.

The commonality between both services is that no dedicated server has to be provided to use the services. This eliminates the entire administration and maintenance effort, so that the resources freed up can be used for other projects and developments.

SaaS solutions include, for example, Google Drive or Slack, which are used by a large number of companies for day-to-day tasks. The scope of functions in SaaS is often clearly defined, so developers can’t create additional functions. In contrast, with FaaS, developers can freely define desired functions according to existing requirements.

• Faster development: Developers don’t need to spend time configuring and managing servers. All available capacity can be used to write application logic and develop applications faster.
• Automatic scalability: FaaS solutions are scalable by their very nature. This eliminates the need for up-front considerations about workloads in terms of data volumes accessed or usage volumes. The required scaling is performed automatically by the provider and is billed on a per-second basis.
• Efficient cost structures: With Function as a Service, you’re only billed for what is actually accessed by users. There are therefore no costs for any resources held in reserve.
• Minimal administrative effort: With FaaS, both the hosting of a server and the management and maintenance of the server infrastructure are not in developers’ hands.
• Freedom in programming: FaaS solutions are usually possible in all common programming languages, so that no additional programming knowledge is required for implementation.

• Limited design options: Since the infrastructure and its management sit with the provider, developers must work with predefined systems. If you want to determine the server infrastructure yourself, but want to save resources for maintenance, it’s worth taking a look at Backend as a Service (BaaS).
• Complex testing: Depending on the provider, transferring FaaS code to a local test environment is not a simple undertaking. Testing an application may require more resources to complete.