Handling 350k Requests for $3 using Lambda

Since its launch towards the end of 2014, Lambda has created a lot of hype, and more importantly, brought the term “serverless” into the industry. At its core, “serverless” does not mean that there are no servers or anything; there are servers, but you are not the one managing those servers. You don’t know where your code is running, you don’t SSH into a server, you don’t do anything that you’d do with regular servers or VMs. You just give a function you’d like to execute to Lambda and let it run the code for you.

As an example, let’s say you have a function that takes a list of products and calculates the price of the whole cart, including various subtotals, shipping prices, and grand totals. If this is one piece of a bigger service, then it might be fine to keep it there. However, if this is a functionality that will be shared between different use-cases, or if this is a full service on its own, then you need to consider how to deploy, manage, and scale it. This exact point is where serverless platforms like Lambda jump in:

• Build a ZIP archive of your code
• Upload it to Lambda
• Map an API Gateway to your lambda function
• Bam, you are done: you have a service that can scale infinitely, requires no maintenance, and can be deployed straight from a ZIP archive.

Alright, the “ infinite scaling ” part may be a bit stretch. It is still software that is running on physical hardware, and there are a bunch of challenges maintaining a smooth scaling from one request to one million requests per minute , but in practical terms, you as the developer don’t need to care about it, the platform itself will handle that.

Devo: My Ideal Usecase

The reason I have been using Lambda is my side project, Devo . Devo is a simple open-source new-tab extension that displays home pages of a bunch of platforms, such as GitHub Trending , Hacker News , Product Hunt or DEV . I found myself visiting these pages quite frequently, and I thought “ why not have them ready in every new tab? ”, which then lead to the project itself.