- It was created with the intention of creating a universally functional, simple-to-configure mechanism for edge services.
PlanetScale serverless driver can undertake intense workloads
Online sports and pop culture news provider Barstool Sports is one of the customers of the PlanetScale database looking to use the serverless driver.
According to Andrew Barba, head of engineering at Barstool Sports, the company faced scaling and performance issues while using Amazon Aurora to operate its MySQL database workloads.
Like PlanetScale, Amazon Aurora is designed to work with MySQL. After the database provider introduced its fully managed database-as-a-service (DBaaS) product in 2021, Barstool switched to PlanetScale.
“Similar to how compute is moving the edge, it’s time for databases to move to the edge as well,” stated Barba.
A developer would have access to a complete set of APIs with a standard application stack that runs on a server, allowing connections to any database.
Van Wiggeren stated that users have had to utilize relatively complex workarounds to connect to various types of databases using edge services, such as Cloudflare Workers. The goal of PlanetScale was to develop an easy-to-configure mechanism for edge services that will operate with any edge service.
The most challenging aspect of developing the serverless driver for PlanetScale was figuring out how to expose every feature of the database through an API that connects via HTTP. It was a process that took PlanetScale around six months to complete as it developed the platform architecture and database capabilities.
“We started out and realized what we actually needed is a new load balancer, a new front door into the database that can serve queries over HTTP. We weren’t willing to compromise on security or functionality, so we actually built and architected from the ground up, our very own HTTP load balancer for MySQL databases,” explained Van Wiggeren.
There are still certain areas in terms of API connectivity that PlanetScale may try to solve in the future. GraphQL, another API strategy for linking and querying data, might be one of these areas.
GraphQL is “an area that we’re really interested in. It’s definitely an attractive and, I think, very interesting way that people are leveraging their databases,” according to Van Wiggeren.