Databases are a key piece of any service or application, because is where gonna live your most precious assets: your data. Since my post “Serverless is Ready, Developers are Not”, I talked with a lot of peers about some issues in the space and a lot gravitates around tools and a great deal is about ways to handle data.
Relational Databases and FaaS don’t play nice together by nature: RDS are connection-based, so they should work with a few long-lived clients, such as web servers, not the fast, ephemeral, spike, nature of serverless workloads.
Even with the great latest improvements like Amazon RDS Proxy and cold starts becomes more and more a UX question (depending, of course, of your runtime and use-case).
Still, RDS is the most familiar pattern to handle data and some could say, the “standard” one—NoSQL as a category, describes tools wildly different, since it encompasses all that are not SQL, all stacked together.
Jeremy Daly in Building Resilient Serverless Systems with “Non-Serverless” Components on Serverless Days 2020 made a great talk about still use RDS models with FaaS. Is not “easy-peasy” or plug and play, as you can see (I highly recommend this video, is also a crash course on FaaS architectures):
So, in this series I’ll experiment with a lot of options available, use-cases, trade-offs and get a hands-on and share my experience here. This is the starting of a series of serverless databases in AWS. I will throw third party offerings in it too, which I can use on AWS. Below is my list. If I’m missing some database or completely getting something wrong—please, let me know!
In the next post in this series I will go over one of these databases, try to use on a use case a little more than “Hello World”—all is shinning and pretty in “Hello World” projects. So, please point me out databases missing, use cases you think I should approach and so on!