This article was originally published on VMBlog: https://vmblog.com/archive/2020/07/02/faas-why-do-companies-have-their-own-serverless-functions.aspx#.XwQbmpMzY1I
The pay-per-use economy of serverless is pushing computing into the commodity space in Wardley Maps slowly but consistently. Unsurprisingly (to me at least), this movement is not pushed solely by cloud vendors, but also by tech giants such as Twilio and Atlassian as they re-package FaaS for the needs of their customers and charge them with the pay-per-use scheme. The last prominent move in this approach came by Salesforce last week, when they introduced serverless functions on their platform. Companies are trying to take advantage of serverless by addressing their customer’s needs. That’s why companies are taking these steps, a move that is completely logical. In this blog post, I’ll go over examples of re:FaaS, discuss the current situation, and where we, the serverless community, are heading with this.
Serverless all the things! Examples of FaaS by tech companies
Of course, Salesforce is not alone in providing ways to take advantage of serverless. There are many different types of customers who need to build their own custom logic on their favorite platform, and there are companies that answer these needs. Let’s check them out!
It’s all about providing a better experience for customers, and doing it right. Twilio achieved a way to take advantage of serverless functions by letting its customers interact with Twilio applications with no need to provision and maintain a server. Instead, Twilio carries all the operational burden of these functions while customers can make the most of the product with minimal effort. Plus, it’s pay-per-user again, so you don’t need to pay anything for servers if you don’t need them, and neither does Twilio. If a company is searching for a solution to send SMS and/or make calls to their customers, Twilio will have a clear competitive advantage just because of this.
JAMStack has created a new way of building fast, fancy, reliable websites based on popular frameworks. Netlify is mostly known for its best ability to serve static content but it also provides a way to run custom logic, such as a form submission or even an API. Guess what? Yes, they accomplish this with serverless functions powered by AWS Lambda. Developers who are accustomed to Lambda can start writing serverless functions right away in Netlify and build any logic on their website.
Similar to Netlify, Vercel also lets developers build super cool websites using JAMStack with what they call server(less)-side rendering. Vercel provides many edge locations to deliver content in a seamless way without recomputing it. This means many companies are now able to build their own websites and, most importantly, statically serve it to their customers at the nearest edge. For Vercel, serverless functions enable developers to embed custom approaches on their websites. Unlike Netlify, Vercel simplifies its serverless function signature by only accepting requests and responses as parameters, which is achieved by wrapping the original AWS Lambda environment and simplifying it only for serving content purposes. Brilliant idea!
Atlassian differs from all of the above with their approach, which enables people to use serverless for building top-notch software. All the other approaches provide a practical way of building custom logic on their platform. Atlassian, on the other hand, seems to be aiming to let developers create secure, reliable applications, and most importantly integrate these with the rest of the Atlassian ecosystem by providing an encapsulated serverless environment named Forge (I’m guessing this is via AWS Lambda). With this approach, they’re literally reselling serverless by adding more value on top of the value that already exists. The most prominent advantage will be streamlining serverless application development into the Atlassian ecosystem by integrating CI/CD security with strong solutions that already exist. Forge is still in beta, and I’m excited to hear some feedback from those who have already tried it.
What are the implications?
It’s not surprising to see the examples of Vercel, Netlify, Twilio, and Salesforce. We’ll see many other platforms take the same steps because you can’t implement every feature request yourself, but you can let customers implement them themselves and charge them for it.
Atlassian’s approach takes this one step further and repackages serverless by strengthening its promises. We can expect to see more platforms that let developers focus on only the value they bring to customers. These platforms are taking the promise (i.e., don’t maintain the scalability and availability) one step beyond and just want people to really write the body of the function with an opinionated essence in their platforms.
It’s a win-win situation for both customers and companies. Companies can still make money on a varying amount of compute that scales down to zero when there’s no traffic. Customers can build their own logic to their favorite platform without having to wait for their feature requests to be implemented.
As a person who’s been a member of the serverless community for years, I’m very glad that there are new channels that let people test the waters by creating serverless functions on their favorite platform. I frankly expect to see many other companies grasp this opportunity to democratize their product with serverless functions, as their customers will be able to build functions according to their needs without requesting permission from the platform. I also expect to see more examples like Atlassian Forge that provide integrated application development platforms that lean on serverless.
With all the improvements, Thundra will be there to support any use cases with serverless just like we are already. Tune in to our upcoming updates by following us on Twitter and subscribe to our updates.