What is Function as a Service (FaaS)?

It is helpful to understand how FaaS compares to other cloud service models like platform as a service (PaaS) and infrastructure as a service (IaaS). While FaaS belongs to the broader "serverless" category—where the provider manages the servers—it offers a distinct level of abstraction. Serverless encompasses any service—such as databases, storage, or compute—where the provider manages the servers. FaaS is specifically the compute part of serverless that runs code in response to events.

• FaaS focuses on event-driven code. You deploy a single function that performs a specific task in response to an event, such as a file upload. The platform handles scaling automatically, often down to zero when idle.
• PaaS provides a platform to build and run entire applications. While PaaS also removes the need to manage hardware and operating systems, it typically keeps your application running continuously, waiting for user requests. It’s best-suited for complete web applications rather than individual event triggers.
• IaaS offers the most control but requires the most management. You rent virtual machines (VMs) and storage, but you are responsible for managing the operating system, runtimes, and middleware.

The FaaS workflow can be simple and efficient. It often relies on an event-driven model where actions trigger your code.

1. Code deployment: You write a small piece of code, or a function, that performs a single task and upload it to your cloud provider.
2. Event triggering: An event occurs, such as a user clicking a button, a file upload to storage, or a message arriving in a queue.
3. Resource allocation: The cloud provider acknowledges the event and instantly provisions the necessary execution environment.
4. Function execution: Your code runs, performs its task, and produces a result, such as processing data or sending a notification.
5. Deactivation and billing: Once the task is complete, the resources are released immediately, and you are billed only for the milliseconds your code ran.

One of the most common issues developers face with FaaS is the "cold start." When your function hasn't run for a while, the cloud provider spins down the environment to save resources. The next time an event triggers that function, there is a slight delay while the system sets up the environment and loads your code. This delay, called a cold start, can slow down your application.

If your application requires instant responses, you can use these steps to minimize latency.

1. Reduce your code size: Review your code and remove any large libraries or dependencies that aren't strictly necessary. The smaller your function, the faster it loads.
2. Use lazy loading: Structure your code so it only loads heavy variables or connections when they are actually needed inside the function logic, rather than loading them immediately when the function starts.
3. Configure minimum instances: Most major cloud providers allow you to set a "minimum instance" count. This keeps at least one instance of your function warm and ready to go at all times, eliminating the startup delay.
4. Choose a faster runtime: Some programming languages, like Go or Python, generally start up faster than others, like Java. If speed is critical, consider writing latency-sensitive functions in a lighter language.