Sail towards lower latencies and higher performance for your NoSQL databases with Scylla
No need to choose between a rock and a hard place! Developers looking for an open source NoSQL database with consistently low latencies and high performance can set sail towards Scylla. Scylla 3.0 is an open source alternative to Apache Cassandra with a whole boatload of new features. Let’s explore!
Choosing the right NoSQL database is difficult. Developers have to weigh issues like performance, speed, consistency, and more. With the release of Scylla 3.0, that choice just got easier.
Scylla is an open source NoSQL database that offers horizontal scale-out, low latencies, and will absolutely not eat six of your sailors as you pass by. Designed from the ground up in C++, Scylla promises to lower the number of database you need as well as optimize dynamic workloads.
Scylla 3.0 comes with a number of exciting production-ready features for NoSQL database developing. Plus, it’s compatible with Apache Cassandra. Let’s take a look at some of these features.
Materialized views are basically denormalization from client-side implementation, except applied to your NoSQL database. Scylla moves this over to the servers, making the implementation faster and more reliable.
Compared to Apache Cassandra, Scylla’s global secondary indexes can scale to any size you need and be ready for production. Essentially, every secondary index creates a materialized view under the hood, using all the columns of the original base table’s primary key and the required indexed columns. These secondary indexes have a smaller footprint and less overhead than materialized views. They also allow for more efficient querying, with paginated, filtered, and customized results.
Speaking of filters, another new feature in Scylla makes it possible to ask more complex queries and get a subset of matching results. Plus, allow filtering reduces the amount of data transferred between the cluster and the application, since it is done on the server. However, some queries might be denied if they affect the performance too much, including ones with non-partition key fields, or parts of primary keys that are not prefixes. That said, filtering allows for compatibility with Cassandra, Spark-Cassandra, and legacy datasets.
Other improvements include a new file format that is designed to reduce storage volume by up to 3x! It’s compatible with existing Cassandra 3.x files while improving speed on previous versions. Hinted handoffs allow for developers to hang onto an unresponsive node and mark a note for when it is ready to work again.
Additionally, Scylla 3.0 offers full, multi-partition scan improvements with improved system responsiveness and lowered disk operations as well as streaming improvements.