Timescale Database Now Available in 76 Cloud Regions
Looking for a time-series database to power applications in a specific part of the world? If so, chances are good that Timescale has you covered with its hosted Timescale Cloud offering, which was recently expanded and now is available in 76 regions across AWS, Azure, and Google Cloud.
Timescale launched Timescale Cloud about a year ago as a fully hosted service for TimescaleDB, the open source time-series database that it also develops. The cloud expansion means that customers in far-flung parts of the world don’t have to compromise when implementing a time-series database to power real-time use cases in IoT, DevOps, IT monitoring, oil and gas, and marketing analytics, says Ajay Kulkarni, CEO of Timescale.
“As a developer, you really want your database to be co-located with your application,” Kulkarni says. “Our customers are in Japan, India, and Australia, markets you typically wouldn’t reach when you’re an early stage company like ours, but we’re reaching them because we’re essentially available in their backyard.”
Timescale is gaining traction with its time-series database, which is built on PostgreSQL. The database adopts the relational data model, and keeps time-series measurements in its own row, while allowing other fields to be added as the user sees fit. The database supports ANSI SQL, enabling developers to get started quickly. The company doesn’t disclose the number of paying customers, but the open source TimescaleDB product has been downloaded tens of millions of times and accounts for more than 500,000 active databases, according to Kulkarni.
The most common use case for the time-series database is serving real-time data to dashboards that provide customers with visibility into data flowing off sensors, IT gear, manufacturing lines, airlines, online music sales, or whatever else the customer wants to measure, Kulkarni says.
For example, Everactive, a Virginia company that manufactures battery-less sensors that run on kinetic energy, has adopted Timescale Cloud to better track data coming off sensors. Schneider Electric, Warner Music, Bosch, Samsung, and IBM are also using TimescaleDB or the company’s commercial offerings.
“As companies undergo digital transformation and they’re better able to track data in real time,” Kulkarni says. “Essentially what that means is you can, in much finer detail, measure what’s happening.”
So what’s fueling the rise of time-series databases? According to Kulkarni, traditional relational databases and newer NoSQL databases are ill-equipped to deal with the volumes of real-time data that customers want to capture, store, and analyze today. That mismatch has given rise to—and driven the growth of–the time-series database niche.
“Time-series is a hard-enough problem where having something that’s purpose-built can lead to a better experience and real efficiency for compute and storage costs,” Kulkarni says. “Typically you have your relational data in one database and time-series in some NoSQL store. With Timescale, you put it together in one database. You’re migrating to Timescale, but it lets you de-silo your data. It lets you combine them in the same place, which simplifies your stack, while also getting better performance and lower cost.”
New compression algorithms introduced with the recent launch of TimescaleDB version 1.5 deliver 15x compression, which translates into a 95x saving in storage costs, Kulkarni says.
The data is “relentless,” he says. “It piles up very quickly. And so traditional database don’t scale for these workloads, which is why you need a database like ours that lets you scale, while maintaining resource efficiency, compute efficiency, and storage efficiency.”
The company is ramping up development of TimescaleDB 2.0 in about a month. The new version will bring support for multi-node deployments, which will help companies with the biggest data demands scale the database horizontally.
Gaining visibility into real-time data via a dashboard is the driving force in many Timescale deployments, because it gives customers something they didn’t have. The next phase will be delivering predictions based on real-time data. We’re not quite there yet, Kulkarni says.
“A lot of companies today want to get there,” he says, “but they’re still laying the foundation, which is you need to first collect the data. You need a place to store and analyze the data, which is where we come in. Once you get that foundation, then I think you can build the layer on top. Right now it’s around real-time monitoring. The next step is more predictive monitoring.”