Serverless Computing

A new architecture that privileges you to not bother about infrastructure, servers, memory usage, scaling of instances, availability, etc including other resources.  We only need to bother about code/business only and it is known as Function as a Service (FaaS).

We can also call it as Server-less Computing.

Serverless means an infrastructure that is made up of servers and we don’t need to manage them. It is a managed offering, we provision these headaches to cloud providers. Pricing model says, don’t pay until nobody is using your application services which means we only to pay for computing, not for deploy.

Now come to more details, Let’s suppose we have a backend service and that it is actually sitting idle.  A request comes from the web / mobile platform to my backend service, instantly servers provisioned and serving the request. As traffic goes up, auto-scaling happens by default. The same with databases as well, it also scales automatically.

Major features of serverless computing are:

  • Event triggers on incoming requests and instances have been provisioned,
  • Auto-scaling of instances, databases,
  • No manual interference, we don’t need to up and down servers, it’s managed by vendors (cloud providers),
  • No incoming request then no compute and no pay.

Some points are needed to be considered before going for serverless computing:

  • Quick Startup: Application must be started or say goes up as quickly as possible comparatively from normal applications. So, we can serve the incoming immediately.
  • Vendor-Specific: Applications are made up of vendor-specific like AWS or GCP. So, we need to choose anyone out of several cloud providers.
  • The application should be Event-based, Stateless.

Note: Later, we add more about serverless using AWS Lambda, Google Cloud Functions.

Leave a Reply

Your email address will not be published. Required fields are marked *