PlanetScale is today unpacking the new Database as a Service based on Vitess, a branch of MySQL distributed horizontally and developed on YouTube. Those who suffer from the consistency and non-schematic nature of NoSQL databases are among the potential new hires targeted by Sam Lambert, Chief Product Officer of PlanetScale, for database services.

Vitess A horizontal database aggregation system for MySQL created by YouTube In 2010, we tracked data related to a particular video, such as views and other metadata. By design, MySQL runs on a single server, Google Affiliate video exceeded the capacity of a single scaling server. The company therefore devised a way to fragment MySQL data on several servers in a cluster.

Over time, other large web companies such as Slack and GitHub have embraced Vitess. According to Lambert, the deployment has expanded, including those that have reached 70,000 nodes in the cluster. Co-founded in 2018 by Jiten Vaidya and Sugu Sougoumarane, the first co-creators of Vitess PlanetScale Our goal is to set up a company centered around the open source Vitess project. That same year, Vitess was accepted as an incubation project at Cloud Native. IT Foundation ..

The company took an important step today by providing new customers with its database as a service from Vitess. Lambert says this offering will allow customers to launch distributed databases in the cloud to enhance any application that plans to communicate with the MySQL backend. Customers can enjoy all of the scalability benefits of Vitess, but without having to manage a distributed database.

According to the former vice president of engineering at GitHub, using databases as a new service is ridiculously straightforward, at least when it comes to setting up and managing traditional distributed databases in the cloud.

“We are not asking you to provision nodes, replicas, etc. We don’t need to provision clusters. We do the supply internally, ”explains Lambert. Datanami.. “We basically believe that we don’t have to ask our customers about this at all. They just want to back up their data and get things done.

The “less, more” mantra extends to the specific cloud provider on which the PlanetScale data runs. Companies in Mountain View, California use cloud services such as: AWS and Google Cloud In order to serve the database as a service, this does not actually give the customer the ability to specify on which cloud the data should reside (although companies with these requests will need PlanetScale as needed. You can use to place the data). ..

“We make sure we’re close to where our users are,” Lambert says. “When provisioning resources on Amazon and Google, this is an implementation detail to provide the best experience.

MySQL has 25 years of experience in powerfully enhancing transactional on-premises and cloud-based applications, and PlanetScale aims to provide robustness for users who need a solution to the dilemma of horizontal scalability. According to Lambert, the offering passed the ACID test and chose data consistency as the primary goal.

The main goal of Vitess is to ensure the integrity of data and transactions, but that does not mean that other departments such as speed compromise. MySQL has proven to be faster on its own Redis MySQL is “surprisingly fast,” Lambert said.

So why aren’t more and more organizations using MySQL or its distributed cousins ​​for large-scale data deployments and applications? According to Lambert, the answer comes down to developer friendliness.

“A fair criticism of MySQL is that it is difficult to use as a developer,” he says. “There is a lot of orchestration and you have to do it. “

Lambert says PlanetScale has been working to make MySQL easier for developers to use, just as the configuration details when running Vitess in the cloud have been simplified. Specifically, he says, a method has been developed to push schema changes to all nodes in a Vitess cluster in an easy-to-use manner.

The fact that Vitess uses schemas to organize data is a major advantage of relational databases and is not destroyed like many NoSQL databases, Lambert said.

Specifically, Lambert pointed out the MongoDBA NoSQL database which is very popular with web and mobile application developers due to its ease of use. Lambert says he’s using MongoDB at another company and finds it wrong to destroy the schema.

“By doing something that isn’t drastically scalable internally, we’ve avoided the difficulty of applying a data schema, which many companies are known to lose data on,” MongoDB Architect said. He says. “I think it’s the world of databases. There are many databases that are poorly rated and have bad tradeoffs. “

With millions of deployments over decades, MySQL is a known quantity. By moving horizontally through the Vitess project, tackling the challenges of schema management, and radically simplifying the configuration of databases in the cloud, the people at PlanetScale will be a great combination of web and mobile applications in the future. . I think I have achieved it.

“That’s the ‘getsh% & done’ stack,” says Lambert. “You can tinker with new technology or choose stable technologies to innovate where your customers expect the product. bells and whistles. Nobody cares to be there. Nobody cares? “

PlanetScale’s database service is in beta. You can register on the company’s website at www.planetscale.com.

Related products:

Large-scale database migration

Cloud Configuration Errors Drive Database “Check In”

Cloud Now’s Default Database Platform, According to Gartner