The open source SycllaDB NoSQL database continues to gain new features and users as it ramps up its plans for 2021.
At the Scylla Summit 2021 virtual event which ran from Jan. 12-14, ScyllaDB CEO Dor Laor shared new features and the roadmap for the NoSQL database's future.
A key part of ScyllaDB's roadmap is Project Circe, a yearlong initiative that aims to bring new performance and consistency to the database. The Summit was also highlighted by multiple users that outlined their ScyllaDB deployments, including Ticketmaster, Expedia Group, Zillow and GE Healthcare.
ScyllaDB NoSQL at Expedia
ScyllaDB is an open source NoSQL database that is positioned as a drop-in replacement for the Apache Cassandra database. It is that use case that attracted Expedia Group to ScyllaDB. Expedia Group is one of the world's largest travel platforms and uses Apache Cassandra for many of its applications.
In a user session on Jan. 13, Expedia developers explained why they moved some of their workloads to ScyllaDB.
Singaram Ragunathan, data architect at Expedia Group, explained during the session that Cassandra, while powerful, comes with challenges. He noted that in his experience it takes a significant amount of time and effort as well as expertise to properly tune Cassandra for optimal performance. In particular, Expedia experienced some issues with performance when there were bursts in traffic. The development team responded by adding buffer nodes to increase speed and capacity, which ultimately resulted in more infrastructure costs.
Expedia tried out ScyllaDB in an attempt to accelerate performance, Ragunathan said, while lowering infrastructure costs. The basic idea was that with the same number of virtual compute instances on the AWS cloud as it used with Cassandra, Expedia could deliver more capacity and throughput by switching to ScyllaDB.
Moving from an Apache Cassandra code base, it's frictionless for developers to switch over to ScyllaDB.
Singaram Ragunathan Data Architect, Expedia Group
"Moving from an Apache Cassandra code base, it's frictionless for developers to switch over to ScyllaDB," Ragunathan said. "For the use cases that we tried there wasn't any data model changes necessary and ScyllaDB was entirely compatible and a drop-in replacement."
With a few small adjustments to the Expedia automation framework that provisions the company's Apache Cassandra clusters, Ragunathan said his team was able to provision ScyllaDB open source clusters with positive results.
"We were able to store more data per node and achieve more throughput per node, thereby saving significant dollars for the company," Ragunathan said.
ScyllaDB Project Circe moves forward
In his opening keynote, ScyllaDB CEO and co-founder Laor highlighted the progress the open source database has made since the last Scylla Summit in 2019. With ScyllaDB 4.0, which was released in May 2020, the project achieved a number of milestones, including feature parity with Apache Cassandra as well as support for Amazon's DynamoDB.
While progress was made in 2020, Laor emphasized that there is lot of work to be done in 2021.
"We'd like to focus on elasticity and maintainability of ScyllaDB over the course of the next year," Laor said. "We believe that even today we're doing a better job … but we still also have rough edges that we like to polish."
To smooth the "rough edges," which include database consistency and performance items, Laor launched Project Circe. The name is drawn from Greek mythology, like the name Scylla itself. Project Circe will provide monthly updates with new development to help improve ScyllaDB elasticity, consistency, stability and performance, as well as ease of deployment, according to Laor.
Among the key area of improvement is consistency through use of the Raft consensus protocol. Currently without Raft, ScyllaDB much like Apache Cassandra, offers the promise of eventually consistent operations.
"With Raft you'll be able to have strong consistency without any performance penalty," Laor said. "So that is outstanding in terms of data consistency."
