DOORDASH SUCCESS STORY UPGRADE
DOORDASH SUCCESS STORY SOFTWARE
DOORDASH SUCCESS STORY CODE

Function blocks that look like bodies.

Already having the capability to do copy and equals simplifies daily chores that would have been some IDE-generated junk or a Lombok hack.

Data classes are something that POJOs or DTOs (Data Transfer Objects) were always missing.

DOORDASH SUCCESS STORY CODE

If used right they make code look simpler and more pleasant to read.

Extension Functions is a lifesaver and an answer to utility classes.

Yes, interoperability with Java might still have some nuances, but as long as you have written your code inside Kotlin itвЂ™s a lifesaver (almost all of our business logic is in Kotlin).вЂќ Thanks to KotlinвЂ™s extensions system, we ended up writing additional extensions to weave these constructs together and still write some delightful code.вЂќ Favourite feature вЂњNullability and safety is our favorite feature due to the guarantees it can put on the code. вЂњWriting interoperable code between libraries can get tricky. There was some work done to glue them together, including adding interceptors, extension methods, and dependency injection, but the ecosystem quickly reached a stable state. These libraries provide the foundation for out-of-the-box coroutine reactor extensions. Building on top of Netty, they were able to add libraries like Lettuce (for Redis), R2DBC (for Postgres), and Micrometer (for instrumentation). In DoorDash, Kroto+ is used to generate the coroutine-friendly client and server code. The common backend stack means that developers can write shared libraries for services and the BFF. Kotlin also allowed the team to use the rich Java library ecosystem and build solid services with complete instrumentation and maximum visibility. In such a system, gRPC servers are capable of writing end-to-end suspendable code without descending into callback hell. The BFF uses Kotlin either with SpringBoot or Micronaut.

DOORDASH SUCCESS STORY SOFTWARE

The DoorDash software backend now consists of two types of services: вЂњbackend for frontendвЂќ (BFF) services that act as stateless proxies, and internal services that run on Kotlin and gRPC and are accessible only through BFF. Kotlin also allowed developers with existing JVM experience to draw on their existing knowledge while building on top of solid frameworks that were well-documented and battle-tested. It was easy to read, understand, and maintain. Kotlin looked promising with its clean syntax and support for coroutines. The JVM appeared to be the most attractive landscape in terms of libraries, tooling, and frameworks, but almost all the developers were concerned about JavaвЂ™s ceremonial constructs and bloated code.

Golang performed great but at that time it wasnвЂ™t suited to complex business logic. The performance achieved with Python wasnвЂ™t sufficient. So DoorDash started to look into the options available for making a more scalable solution.ДЄ special internal committee evaluated the choices available, such as Python 3 (with async IO), Java, Golang, and Kotlin in combination with gRPC to see which would best fit the companyвЂ™s needs. вЂњWe are serving millions of users every day, delivering their orders with a very intricate, well-knitted solution.Our Kotlin services are handling requests from several complex logistic systems in order to weave together a seamless on-demand food delivery experience for the customers.вЂќ Why Kotlin?ДoorDash had a Python monolith system backend that worked satisfactorily but wasnвЂ™t going to keep up with the projected scaling. San Francisco-based on demand logistics platform DoorDash deals with customers who order food, merchants who prepare food for customers, and вЂњDashersвЂќ who deliver the food to the customerвЂ™s door.Д«y the end of November 2020, DoorDash had achieved $1.9 billion in revenue and itвЂ™s one of the most anticipated IPOs in 2020.

DOORDASH SUCCESS STORY UPGRADE

This upgrade resulted in tremendous improvements:ДЄnother great thing about Kotlin is that people coming from Java, Golang, and Python find it really easy to get started. Two years ago, they made the bold move to split their existing legacy Python monolith into multiple backend services, which they developed in Kotlin. Hundreds of engineers work with Kotlin across a massive system spanning multiple repos. DoorDash в†— uses Kotlin for all their backend services and infrastructure.

Doordash success story