ScyllaDB has introduced the discharge of model 6.0 of its database, which includes a new replication structure referred to as “tablets.”
The brand new structure builds on the Raft consensus protocol and supplies enhancements to elasticity, pace, simplicity, and effectivity. ScyllaDB claims that now clusters will be doubled inside quarter-hour, and their throughput in that very same time interval is quadruple what it was.
“With ScyllaDB’s new tablets structure, information is dynamically redistributed because the workload and topology evolve. New nodes will be spun up in parallel and begin adapting to the load in close to real-time. This implies groups can shortly scale out in response to site visitors spikes – satisfying latency SLAs while not having to overprovision ‘simply in case,’” the corporate defined in its announcement.
In keeping with the corporate, tablets are perfect for eventualities the place there will likely be various site visitors quantity all through the day, sudden spikes, or main occasions that can improve site visitors to an unknown quantity, reminiscent of sporting occasions or Black Friday.
The corporate will likely be sharing further particulars on the brand new structure and its efficiency enhancements later this yr at its P99 Conf on October 23-24.
“We used to outline ScyllaDB as the ability of Cassandra on the pace of Redis,” mentioned Dor Laor, co-founder and CEO of ScyllaDB. “With the ScyllaDB 6.0 launch, we add a 3rd pillar to the equation: the usability of DynamoDB.
You may additionally like…
Q&A: How cognitive fatigue impacts developer productiveness
4 inquiries to reply earlier than scaling a product
