DataStax is making serverless, NoSQL database-as-a-service (DaaS) Astra DB, built on open-source Apache Cassandra, available in multiple regions across major cloud services providers such as Microsoft Azure, Amazon Web Services (AWS), and the Google Cloud Platform (GCP).
The announcement follows rival MongoDB’s launch in July of its own serverless, NoSQL DaaS, Atlas. Enterprises have been looking to adopt serverless architecture as it eases the cost of owning and maintaining servers. In a serverless setup, companies only pay for the functions they use — mainly, storage and compute.
Making Astra DB available across cloud providers and in different regions could be a game-changing strategy for the company, said Craig Kitterman, DataStax’s vice president of product management in the database division.
Multiregion availability cuts complexity
In the past if any company wanted to have a multiregion setup in order to do data replication and ensure real-time availability of their data, they would have to embark on a complex infrastructure configuration journey that would be expensive given challenges such as network load balancing and security along with the whole plan of scaling up and down, Kitterman said.
“What we've done is we have distilled that entire suite of complexity down into a couple of clicks and a couple of minutes," Kitterman said. "As a result, the cost performance is phenomenal, because customers basically don't have to employ all those experts and spend all that time and energy and months of effort, just to try to get it right. We've already got it right.”
A multiregion database strategy is important for geographically dispersed enterprises, analysts said.
“For businesses that have data that must be shared across regions in real time, multiregion is critical. The alternative is database replication, which introduces considerable latency,” said Carl Olofson, research vice president of data management software at IDC.
Multicloud availabiity gives customers choice
With the move to go multiregion on a range of public clouds, DataStax has opened up choices for customers, said Doug Henschen, vice president and principal analyst at Constellation Research.
“Many large enterprises use multiple cloud providers and want flexibility. In some cases, one public cloud has better coverage in certain countries and regions than others, so the ability to run a database service across regions and across cloud providers opens up choices that customers might want to exploit,” Henschen said.
In terms of its multiregion, multicloud features, DataStax’s Astra DB service does not have much direct competition, Olofson said. For example, while Google Cloud Spanner is serverless and multiregion, it is not multicloud and only runs on GCP. On the other hand, MongoDB is not multiregion even though it is serverless and multicloud, he said.
Meanwhile, DataStax rivals such as Snowflake, MongoDB, MariaDB and TigerGraph follow a multicloud and hybrid-cloud strategy, but they are not necessarily direct DataStax rivals, Henschen said.
“Most independent NoSQL and relational database management system vendors are differentiating from single-cloud-native services and are responding to customer demand by providing hybrid and multicloud support," Henschen said. "None of these vendors address the same use cases or compete directly with DataStax, but they are all independents following a multi-cloud strategy.”
DataStax distinguishes itself as it is often used for extremely large data stores comprising terabytes or even petabytes of data, and uses a denormalised schema database optimisation technique.
DataStax offers DaaS anywhere, anytime
Any customer that signs up for AstraDB can deploy a database anywhere in the world by just signing up and choosing their preferred choice of cloud provider and region, Kitterman said.
Existing customers can also add a new region through the Astra DB console, which also keeps stock of their compute and storage functions.
“If you already have your database configured, and running somewhere else in the world, you can simply now go in and add a new region. So, in this case, you would maybe add a region in India, and within a few minutes, that database would be provisioned for you and replication would be configured and all your data would be automatically replicated to the new region,” Kitterman said. Once replication is complete, any new rights in the old database would be replicated henceforth in the new region and vice versa.
DataStax also said that it was adding regions on request from customers, where DataStax is not available now, and Kitterman said that most of the requested countries were covered.