Episode #57: Building Serverless Applications using Webiny with Sven Al Hamad

About Sven Al Hamad

Sven Al Hamad is co-founder and CEO of Webiny Serverless CMS. Sven has worked with the largest media and ecommerce customers in Europe as their trusted advisor on the topics of web performance and architecture, and has a proven track record of successful delivery of several multi-million dollar projects for large enterprises. Sven is also an experienced entrepreneur who has acted as a CTO in 4 different startups.

• Twitter: twitter.com/svenalhamad
• Email: [email protected]
• Webiny: webiny.com
• Webiny Twitter: twitter.com/WebinyPlatform
• Github: github.com/webiny

Watch this episode on YouTube: https://youtu.be/9TSmOcLBr0k

Transcript

Jeremy: Hi everyone. I'm Jeremy Daly and this is Serverless Chats. Today I'm speaking with Sven Al Hamad. Hey Sven. Thanks for joining me.

Sven: Hey, Jeremy. Thanks for having me.

Jeremy: So you are the CEO and co-founder of Webiny. So why don't you tell the listeners a little bit about your background and what Webiny is?

Sven: Yeah. So well, in terms of my background, I'm a developer. I started maybe 20 years ago, even more coding websites and many other stuff along the way, but also worked in the enterprise world for several years and decided to start Webiny about a year and a half ago, maybe two years back. But now I'm more focused on the business side. And what Webiny is, it's essentially an open source framework for building full stack applications that deploy to serverless infrastructure like AWS Lambda and similar. So it's all about creating serverless solutions.

Jeremy: Awesome. All right. So that's what I want to talk to you about obviously the Webiny platform, what you can do with it. And I'd like to start by kind of going through more of the details, right? Because I think we get confused with maybe what a CMS is versus what a application platform is versus what a cloud provider is. I mean, so I think it can get very confusing if you don't dive deep into the docks and even I was looking at the Webiny site and I was like, "All right, it's a CMS, but it also has a framework for building things." And so I'd love to go through that. And then we can talk about a couple of other things just to get your insight on serverless, but let's start at the beginning. So why did you build Webiny? What was the thing that triggered that?

Sven: So I started researching the serverless market in general. That was late 2017 and the more I dived deeper into the potential of serverless and serverless infrastructure, I kind of understood that serverless has such a big potential that actually it can become the standard, how we are building all applications in the future. Essentially, if you want to build an application five years from now, you're going to build it on serverless infrastructure. Serverless is going to become that standard.

And at the same time, I looked at the market, in terms of the solutions that are available today to help you do that. Well, there was nothing that I could use out of the box to help me build an application. There are tools to help you monitor and deploy serverless applications, but there are no tools to help you build actually a full stack serverless application. So I saw a big opportunity there, but also at the same time, I had a web design development agency many years back where we were all testing different CMSs and different solutions on building things. So from that learning and that experience, I decided, "Okay, let's take that. Let's take the serverless market, which is new and has great potential. Let's build a solution for that market that also is open source at the same time, so it benefits the whole ecosystem and the community as a whole."

Jeremy: Yeah. And who among us, hasn't owned a web development company in the past? I know I did for about 12 years and honestly it's funny because we built a CMS. I mean, CMSs were, this was before WordPress, right? So I mean, WordPress comes along and it changes a bunch of things. But one thing that was always a pain for me, and I know we can get into this as part of what the tool offers. But it was things like building forms. If I had to build one more HTML form in my web development company, I mean, I was ready to just, I don't know, go stick my head in a closet or something like that because it was just, it's so tedious, it's so repetitive. And as one of the things I think we've done really well or we've done a good job of is we've abstracted away a lot of these things and tried to make it so that where we make the undifferentiated heavy lifting, much easier.

But a big part of that, and this is something that always scared me, every time you build a new project, it's less about the interface, it's less about, maybe the backend, it's a lot about the data, right? We want to make sure that our data is secure, that our data is backed up. So I'd love to just talk about the data model that you have built into Webiny, because again, it supports a bunch of different things. But could you explain that?

Sven: Yeah, so well just handling data in a serverless environment comes with its own challenges. We found that really early on. So we decided to go with a MongoDB, particularly MongoDB Atlas to store the data, but how you actually talk to a MongoDB database from a Lambda function it's not the best today. If you don't use some specialized solutions, but you also find a similar problem with MySQL. That's why you created the MySQL library to help you do that. But essentially, what we did with Webiny, we have this notion of multicloud in our minds and we didn't want to get locked into specific databases. And we built a data library, a library that handles how we talk to databases and how we model the data models actually inside Webiny.

And that library is also open source, it's called Komodo and through Komodo, we built pretty much all the data models that you see in Webiny today in all our applications. And the beauty is that with Komodo, on the other side, you have these adapters for MongoDB. We also have an adapter for MySQL that's not published there, but we're planning on also building an adapter for DynamoDB and things like that. So we had to kind of also put some constraints on the data model. We didn't want to make it fully no SQL because we also want to support SQL in the future. So we put some constraints there but Komodo is kind of lifting off all the complexities there for us as a user. And on the other side, what we found is another challenge with handling pretty much TCP connections to the database. So we built, if you look at our architecture, we built something called the database proxy, essentially a Lambda function to which all our Lambda functions, talk to.

And only that Lambda function has the actual connections to the data, but it's like a funnel. So you can kind of limit how many connections you send to the actual database, reducing the number of zombie connections and so on. Some database have really, really smart abilities. So you can programmatically tell it, "Kill this connection or open another connection." But MongoDB doesn't have that. So it's pretty much kind of a, just having that funnel really low and then queuing up all the requests there. So but yeah, the data model has being a challenging topic for us there was a lot of iteration, a lot ...