Serverless Architectures are an evolutionary step in software development. They enable developers to express their creativity in meeting user needs, freed from the burden of undifferentiated heavy lifting required to build scalable secure applications. ServerlessConf is about the community sharing challenges and best practices in Serverless Architecture.
These new architectures, are anything but new. They build on the work done in many areas, from Backend as a Service providers such as Firebase & Parse, or the advent of Single Page Apps through to frameworks like AngularJS & EmberJS, to the trend to standardise on RESTful APIs, and cloud functions like AWS Lambda, to name a few of the technologies that have enabled the current state.
Serverless Architectures are the culmination of years of evolutionary development, that have finally led us to a point where full end to end applications can be built by purely using abstracted platforms, without the reliance on an operating system, never mind a server.
Whilst a lot of the foundational technologies that enable these architectures are not new, it was the release of AWS’s API Gateway service that has been a catalyst to enable the growing wave of serverless architectures, as I wrote about at the time. Whilst AWS’s Lambda was proving popular, it was tightly coupled to AWS’s services and infrastructure, by allowing functions to be invoked through a RESTful interface allowed it to integrate with services outside of AWS.
That small change allowed Lambda, through API Gateway, to act as “API glue”. A way to tie disparate abstracted services together in a private secure manner, and apply logic to the integration. This model has recently been validated with Google announcing Google Cloud Functions, and IBM announcing it’s OpenWhisk project, both which mimic AWS’s Lambda, with the key exception that they have a RESTful interface built in from day one. The big players aren’t the only ones with a platform, with Auth0’s webtask.io platform offering a similar “cloud function” through a RESTful interface, available in production.
With the adoption of these architectures now rising, there are three key use cases appearing, fully fledged applications like Property Tour Pro and acloud.guru, data pipelines and configuration/automation like backup handling.
With these new architectural patterns, there are new challenges. There is a lack of tooling around these services, with a multitude of open source projects popping up to solve the problem. What does monitoring even look like for these kind of services, when there is nowhere to install an agent? Now that we’ve distributed our dependancies and use tiny uncoupled functions to integrate it all, how do we manage all of this?
ServerlessConf is a community led conference in New York at the end of May, that is an attempt to explore the full breadth of what Serverless Architectures really are, the problems they solve, and the new set of challenges they bring.