Follow Datanami:
April 9, 2019

ScyllaDB Rolls Cloud Database Service

Databases continue to migrate to the cloud with mix results, often requiring customers to choose between the convenience of cloud services and the countervailing costs associated with scaling and performance metrics like low latency.

NoSQL database vendor ScyllaDB claims to have addressed those issues with a new cloud service running on Amazon Web Services (NASDAQ: AMZN) designed to push real-time applications into production. The six-year-old startup claims its Scylla platform is the fastest column-store database, combining the functionality of Apache Cassandra with the speed of a light key-value store.

The cloud database as a managed service released this week is aimed at deployment and scaling of applications without the administrative overhead. ScyllaDB is willing to match its new cloud database service against other NoSQL platforms such as Amazon’s DynamoDB key-value and document database or Google Cloud’s BigTable NoSQL wide-column service.

One reason is the underlying hardware, which is touted as reducing costs by requiring fewer servers to run real-time and other advanced workloads, the database vendor said Tuesday (April 9).

Scylla Cloud runs on AWS, but the vendor said it outperforms DynamoDB in terms of costs thanks to ScyllaDB’s C++ codebase and “close-to-the-metal” architecture, which leverages multicore processors.

“Existing cloud databases are prohibitively expensive, requiring wasteful overprovisioning, and all too often they can’t meet the [service-level agreements] of today’s real-time applications,” said Dor Laor, CEO and co-founder of ScyllaDB.

Laor also notes that many cloud database services are proprietary while ScyllaDB’s roots are in open source. Laor helped develop Linux distributor Red Hat’s (NYSE: RHT) KVM hypervisor before launching a startup that evolved into ScyllaDB. Prior to its cloud database foray, the ScyllaDB focus on developing a NoSQL database built on its version of the Cassandra open-source database.

While its real-time database is fully compatible with Cassandra, the vendor asserts that its “share-nothing” approach boosts throughput and storage capacity by as much as 10 times over Apache Cassandra.

Meanwhile, its new cloud service allows Scylla users to run their clusters on-premises or in the public cloud, while database administration can be offloaded to the new managed cloud service. That means a developer cluster can be spun up more cheaply, along with the ability to launch production clusters consisting of larger cloud instances across multiple regions.

ScyllaDB said its new cloud database service is available in AWS regions in the U.S., Europe and Asia. It plans to “soon” offer Scylla cloud on Google Cloud Platform and Microsoft Azure.

Recent items:

Scylla Eyes Cassandra’s NoSQL Workloads

Which Programming Language is Best for Big Data?