The developers of the ScyllaDB DBMS have announced the transfer of the project to a proprietary license that limits the scope of application. The ScyllaDB 6.2.x branch will be the last one available under the AGPL license. If desired, interested parties can create a fork and continue developing the ScyllaDB code base under the old license.
Previously, the ScyllaDB project adhered to the Open Core development model, in which the basic part of the product is developed under a free license, and the extended functionality is closed and supplied only to holders of a commercial license. In accordance with the new licensing scheme, the publication of the open edition of ScyllaDB OSS will be discontinued, and the previously closed commercial product ScyllaDB Enterprise will begin to be distributed with the provision of source code under a proprietary license. Instead of ScyllaDB OSS, users will be able to use the full version of ScyllaDB Enterprise for free, subject to certain conditions.
The new license prohibits the use of the software to compete with products and services of ScyllaDB developers, as well as to create cloud systems "application as a service" (SaaS, software-as-a-service) and commercial systems "database as a service" (dBaaS, database-as-as-service). The license does not limit the use of the free product for commercial purposes and in production environments, but on the condition that the total storage size on all cluster nodes does not exceed 10 TB, and no more than 50 VCPUs are used for data processing. Users who do not meet the stated criteria are offered to purchase a commercial license.
For example, you can use ScyllaDB for free on a three-node cluster, if each node contains 16 logical CPU cores and 3TB of disk storage. Such a configuration can process 100-200 thousand operations per second. Distribution of ScyllaDB and modifications to the code are permitted, provided that it is provided under the same license, a full list of all changes is listed, and information about the original author of the product is provided.
The reason for changing the ScyllaDB distribution policy was the desire to unify and optimize the development of competing commercial and open source versions, the separate maintenance of which created difficulties, took a lot of effort and wasted resources. At the same time, due to the complexity of the internal architecture, the ScyllaDB code base was developed exclusively by the company's employees and the open source project never had any third-party participants who contributed their changes. The turning point was the work on implementing the Raft consensus algorithm, tables and built-in support for the S3 API, as a result of which many auxiliary functions were transferred from external applications to the main DBMS.
Ultimately, it was decided to combine the open and commercial products, which on the one hand will simplify development, and on the other hand, will expand the capabilities of the free version. For example, users who fit into the free limits will have access to such features as LDAP support, PGO optimization (reducing text delays by 33% and increasing productivity by up to 50%), file-level streaming mode (accelerating the addition/removal of nodes up to 30 times), setting your own priorities for different loads, support for compression of RPC traffic between nodes using the ZSTD algorithm, an improved data packaging strategy (reducing the storage size by up to 35%), encryption support, a Kubernetes operator for ScyllaDB, a long release support cycle.
ScyllaDB allows you to create distributed, scalable and fault-tolerant noSQL systems that store data in the form of associative arrays (hashes) with several nesting levels. For structured queries, the SQL-like language CQL (Cassandra Query Language) can be used. ScyllaDB-based clusters provide a linear level of scalability, where performance is directly dependent on the number of processor cores. Data placed in the DB is automatically replicated to several nodes, and if a node fails, its functions are picked up on the fly by other nodes. Adding, updating and deleting nodes in the cluster is performed without stopping work and without reconfiguring other nodes.
ScyllaDB was founded ten years ago by Avi Kiviti and Dor Laor, who also created the hypervisor KVM and the OSv operating system. The project was created as an attempt to rewrite Apache Cassandra from Java to C++ to achieve higher performance. The DBMS is API-compatible with Apache Cassandra and Amazon DynamoDB. In tests, ScyllaDB demonstrates a 2-5x throughput increase compared to Apache Cassandra. It is noted that an Apache Cassandra-based cluster can be replaced with a ScyllaDB cluster, which contains 10 times fewer nodes, but despite the smaller number of nodes, it still outperforms Apache Cassandra by 42%.
As for other ScyllaDB products, the Seastar asynchronous server application development framework, drivers, and Kubernetes operator will continue to be licensed under Apache 2.0. The Scylla Manager cluster management platform, previously distributed under a proprietary license, has been transferred to the Apache 2.0 license. The closed implementation of the geographically distributed Kubernetes operator has been merged with the open Kubernetes operator and will be licensed under Apache 2.0.
Source: opennet.ru
