We are excited to bring Transform 2022 back in-person July 19 and virtually July 20 – 28. Join AI and data leaders for insightful talks and exciting networking opportunities. Register today!
Today, PlanetScale, the database-as-a-service company targeting the largest data collections, announced a new collection of features to improve performance for many more complex forms of analysis while simplifying life for developers who must provide it. The company wants to open up the database so that developers can track individual queries and integrate the tool with other services.
PlanetScale is one of several companies like Amazon, Google, Yugabyte, Fly.io and OtterTune that aim to take popular open-source databases and build products around them. Their service is built around MySQL and Vitesse, two very popular open-source options that power many websites big and small.
Disrupting the database market
The marketplace is growing increasingly competitive as more companies announce services. Recently, Google offered AlloyDB, a sophisticated version of PostgreSQL that has been enhanced with Google’s cloud-native data storage experience and artificial intelligence (AI)technology.
PlanetScale distinguished itself in the past by offering a sophisticated and feature-rich way for developers to work with datasets. In the past, databases have offered only one version of the dataset. PlanetScale’s core service allows developers to set up different branches that can evolve separately until the time comes to merge them. This provides developers with a way to experiment with new formats, making it simpler to create new versions. PlanetScale Rewind allows developers to reverse all the changes made to the schema.
The three features in today’s announcement also target the developers. All are said to be in use with some select existing customers and are now ready for everyone.
The first major option called “Insights” offers a dashboard that tracks all the queries and response times. If problems emerge, developers can search through the historical record for the problematic query and rewrite it in the future.
Fully integrated database features
PlanetScale’s service is competing with some other products that track the performance of web applications. Companies like DataDog, New Relic, Dynatrace and Splunk are just some of the well-known brands that help developers understand what is working and failing in their web stack.
Many of these third-party tools also track many moving parts, not just the database. They watch system load and tracknetwork failures, IO backups and more. Still, the database is often the source of the most mysterious bottlenecks. PlanetScale is hoping that the value in the telemetry captured by its Insight dashboard may be good enough to save their customers the need to also buy one of these other services.
“[The other services] are all very cool. But [they’re] a third-party bolt-on solution that is not completely integrated into the database and thus suboptimal,” explained Sam Lambert PlanetScale’sCEO. “We fully integrate and that’s something that means we can produce a unified experience that’s rich with data and allows people to really understand their databases.”
Adding this feature inside the database eliminates any potential glitches or miscommunication that can happen with third-party software.
“We can see every query. We can just find you the most expensive,” said Lambert. “We can say this exact query costs this much money to run. We can really start to show people true optimizations from the database that is doing the informing here “
A database with better connections and fast queries
PlanetScale is also opening up their database for better connections to outside services. This second new feature called “Connect” sends a constant stream of database updates to other applications. The company imagines that others might use it to create services that provide more real-time analytics of the data. In essence, PlanetScale is carving out the job of storing massive amounts for their business and then making it easier to integrate with other products that analyze the information.
“We’re not gonna try and be like these hybrid solutions offering snake oil.” said Lambert, referring to some other companies that want to offer good performances for both retrieval and analysis. “We want to give you the best way of sending your data straight out of your main database to your analytics database.”
Planet Scaleadded a second storage layer, a column-oriented store, to speed up queries that perform analysis. The company estimates that this extra hybrid option may speed up these queries common in generating reports by as much as 100 times.
The third part of PlanetScale’s announcement is a mechanism to set up read-only versions of the database in other regions around the world. This can simplify development and also increase usability for some use cases where information must be shared over long distances.
Lambert used the announcement as a chance to highlight some of the company’s technical flexibility as well. The core product helps developers create versions or branches of databases and then merge them later. Developers who enjoy using tools like GitHub to create forks or branches of their code can now enjoy the same flexibility when working with the data set itself.
“Someone said to me, ‘I can’t believe how quickly you build stuff. How do you do it?’ And actually genuinely one of the answers is because we’re building on top of PlanetScale,” Lambert claimed. “We’re now starting to get this compound effect of the product making is actually faster to build more of the product.”