ScyllaDB has announced the general availability of Scylla Operator to simplify deployment of ScyllaDB on Kubernetes clusters. ScyllaDB is billed as an alternative to the open source Cassandra database on Kubernetes clusters.
In addition, the company has launched the Project Circe initiative, through which it will provide a series of updates to the NoSQL database to improve elasticity and consistency over the next 12 months. At the core of those efforts is a Raft consensus protocol, which will both enable the database to double its size in a single operation and improve transaction consistency. Other Project Circe updates will improve throughput, reduce latency and simplify the operational complexity of running a distributed database at scale.
Finally, ScyllaDB has also made a beta edition of the database available on the Google Cloud Platform (GCP).
Dor Laor, CEO of ScyllaDB, says as organizations build and deploy microservices-based applications, the amount of data that will need to be accessed, per node, on a Kubernetes cluster will significantly expand. ScyllaDB can support as much as 60TB of data per node. ScyllaDB relies on a shared design on each node that places a different subset of data on each processor core. Those cores only communicate when explicitly required by an application. That approach makes it possible to manage petabytes of data in a way that horizontally scales, Laor says.
The ScyllaDB was created to provide IT organizations with a faster alternative to the Cassandra database. Databases such as Cassandra and ScyllaDB are widely employed to process transactions based on a wide column store rather than the traditional row-based architecture employed in a relational database. ScyllaDB is written in C++ rather than Java, but supports all the same primary application programming interfaces (APIs) as Cassandra.
While developers are exercising more influence over the selection of databases as part of a larger DevOps team, Laor says he doesn’t see the need for database administrators abating any time soon. However, as databases become more distributed across IT environments, the nature of DBAs’ roles is changing, says Laor.
The sheer number of databases will continue to require dedicated management expertise, even as those databases increasingly become self-healing, Laor says.
Organizations currently using ScyllaDB include Comcast, Discord, Expedia, Grab, Hotstar, Medium, Starbucks, Ola Cabs, Samsung, IBM and Investing.com. As digital transformation initiatives mature and more applications are run on Kubernetes clusters, ScyllaDB’s scale capabilities will see greater demand; what remains to be seen is at what rate.
In the meantime, IT organizations should consider how to manage highly distributed IT environments that include petabytes of data that must be processed and analyzed in near real-time. That decision must take into account not just which database is best-suited to the task, but also which is simpler to manage at scale.