Compare ScyllaDB Options

ScyllaDB is based on Apache Cassandra (based on Amazon Dynamo paper), it’s a wide column NoSQL database architecture, using partition and clustering (ordering) keys. ScyllaDB is a configurable and scalable ring architecture deployed as multi-node clusters in a single rack/AZ, across racks/AZs, and regions. Partitions are mapped to shards and distributed automatically. Learn more about the architecture here.

ScyllaDB is a drop-in replacement for Apache Cassandra, supporting the same CQL interface and queries.

Asynchronous Operations are executed using Seastar, a standalone Apache library developed by ScyllaDB. Seastar is an event-driven, non-blocking C++ framework with a shared-nothing model that supports getting the most out of every CPU core, associated memory, and high-performance networking.

ScyllaDB is written entirely in C++20/23, implementing the best compiler techniques to maximize CPU utilization. Scylla automatically configures your network interrupts to balance IRQ processing across your cores. Scylla explicitly chooses to read data from the drive ahead of time when it expects follow-up-on-disk access instead of blindly relying on the disk.

Unlike other comparable NoSQL DBaaS offerings, ScyllaDB fully optimizes use of multi-core CPUs and NUMA servers. Learn how we run millions of operations per second at sub-msec average latencies.

ScyllaDB allocates and pins memory to single-threaded shards on each node to optimize memory use in its Memtable and Row-Based Cache, giving it the speed of memory caches like Redis – but parallelized and free of contentious lock coordination, resulting in higher throughput.

ScyllaDB 2023.1 and later support Strongly Consistent Schema Management using the RAFT consensus algorithm. Users can now update the schema concurrently from different regions and as often as they want without the risk of scheme inconsistency. Learn more.

ScyllaDB provides configurable, tunable data consistency. Data written to any node will eventually be dynamically sharded, with each shard replicated based on the replication factor and distributed across the cluster – all without developer intervention or app awareness required.

ScyllaDB Enterprise and ScyllaDB Cloud have further optimizations, ScyllaDB Tachyon, that can improve cluster performance by up to 50% while reducing latencies by up to 35% over its Open Source counterpart.

ScyllaDB Manager automates recurrent tasks, such as backups and repairs. ScyllaDB Open Source users are limited to 5 managed nodes, and Enterprise and DBaaS users have unlimited nodes. Learn more.

ScyllaDB Monitoring is a bundle of four components based on de facto industry-standard open-source systems (Prometheus metric collection, Alert manager, Grafana dashboards, and Grafana Loki log aggregation) that can be deployed as containers or directly onto a host. It also collects aggregated NoSQL performance metrics, logs, and events through ScyllaDB Manager.

ScyllaDB uses Prometheus open source monitoring and an SSL proxy to extract and send all the details of operation seen by the ScyllaDB Monitoring Stack to other 3rd party Observability services (e.g. Datadog, AppDynamics, New Relic, and many others).

 ScyllaDB Operator is an open-source project that gives ScyllaDB Open Source and ScyllaDB Enterprise users an easy way to run and manage ScyllaDB via Kubernetes.

ScyllaDB’s unique compaction strategy means up to 37% savings on disk space over other compaction strategies, lowering operating costs. Learn more.

ScyllaDB supports assigning shares of computing resources to different workloads, ensuring no single process can hold the rest of your database hostage. Learn more.

Fully managed SaaS offering by ScyllaDB on AWS and GCP with the option to Bring Your Own Account (BYOA); see the quick start guide here.

Easily review the status of your nodes, read and write latencies, throughput and storage, historical trends, memory, compactions, alerts about potential problems, and much more. You can drill down from high-level to detailed dashboards and display information by DC, node, or shards.

Production ScyllaDB Enterprise clusters can be set up with just three clicks. Then, scaling both up and out enables you to optimize your cluster configuration to meet specific requirements for both performance and cost.

ScyllaDB Cloud runs a full daily backup of your databases. It backs up data to external storage – Cloud Storage for GCP or S3 for AWS, to a bucket in the same region as the node. If you use ScyllaDB Cloud in your AWS account (BYOA), the backups are located under your user account. By default, backups are retained for 14 days.

Encryption in Transit (Node to Node and Client to Node) using TLS certifications and FIPS tolerant compilation and execution on key OSes. Options are available for nodal communication encryption within a single rack/zone, across racks/zones, and data centers/regions.

ScyllaDB Encryption at Rest (EaR) protects sensitive data residing on your cluster.  ScyllaDB EaR integrates with external key management services stored locally or externally through Amazon KMS or KMIP services.  Transparent Data Encryption (TDE) adds a way to define EaR parameters at per cluster and table levels.

 Manage, authorize, and authenticate user privileges via an LDAP server to preserve database control, security, and risk avoidance. 

Auditing allows the administrator to monitor activities on a ScyllaDB cluster, including queries and data changes. The information is stored in a Syslog or a ScyllaDB table.

 SAML-based Single Sign-On, see the documentation here. 

BYOK allows our customers to proactively prevent and reduce breach exposures, gives them more control over their data and key management policies, including permissions, life cycle, and key rotation rate.  ScyllaDB Cloud supports both AWS and GCP key management systems as master keys for ScyllaDB clusters.  For BYOA accounts, the customer managed keys (CMKs) are used.

ScyllaDB Cloud undergoes independent third-party audits to confirm that it adheres to strict industry standards for security, availability, processing integrity, confidentiality, and privacy.  Current certifications also include ISO 27001, 27017, 27018, and 27000 to support ScyllaDB’s SOC 2 compliance.

Licensing – ScyllaDB Open Source code is available on Github and covered by the Affero General Public License.  ScyllaDB Enterprise and Cloud DBaaS are covered Proprietary software license agreements and ScyllaDB Cloud terms of services and use policies can be found in the policies and agreements section on ScyllaDB.com.

The ScyllaDB Open Source Community is continually growing and highly active. As questions arise, connect with fellow users on the ScyllaDB Community Forum and Slack, join mailing lists, leverage our extensive documentation, and learn for free at ScyllaDB University.

Available for paid licenses and subscription offerings. See the support tier table for differences between tiers.  Note that you cannot unbundle DBaaS tiers and their service support tiers. 

DBaaS offerings continually update and use ScyllaDB Enterprise annual production releases, phased out of general availability after the third subsequent release.  Open Source does not have long-term release support.

24×7 Mission Critical Support is available for paid licenses and subscription offerings. See the support tier table for differences between tiers.

ScyllaDB SRE personnel will provide proactive performance optimization and customized alerts against several key metrics to ensure proper operation and recommend improvements.

Bug fixes, patches, hotfixes, and topics that require communication with and/or escalations to the product engineering teams

Installation and configuration of the ScyllaDB cluster and the OS, utilizing ScyllaDB’s best practices and the expertise of the ScyllaDB team. 24*7 proactive monitoring, including automatic alerts. Rolling upgrades to the latest ScyllaDB Enterprise version – with no downtime.

ScyllaDB cloud uses Apache Cassandra Query language and data model BUT implemented in C++ and optimized for ScyllaDB’s close-to-the-metal design and operation.

Separate ScyllaDB API compatible with Amazon DynamoDB allows you to run your DynamoDB apps unchanged on any cloud or on-premises.

Lightweight Transactions (LWT) allow conditional database updates. ScyllaDB’s LWT implementation is faster and more efficient than Cassandra. Learn more.

Global Secondary indexes are a mechanism in ScyllaDB that allows efficient searches on non-partition keys by creating an index. They are indexes created on columns other than the primary partition key, where each secondary index indexes one specific column. A secondary index can index a column used in the partition key in the case of a composite partition key. Local Secondary Indexes enhance Global Secondary Indexes, allowing ScyllaDB to optimize workloads where the partition key of the base table and the index are the same.

Materialized Views (MV) are a global index. When a new MV is declared, a new table is created and distributed to the different nodes using the standard table distribution mechanisms. It’s scalable, just like normal tables. It is populated by a query running against the base table. It’s not possible to directly update an MV; it’s updated when the base table is updated.

All ScyllaDB drivers are modern, feature-rich and shard-aware, each supporting a different key programming language, including C++, Java, Python, Go, and Rust.

CDC is a production-ready option to track and log all changes made to any table in any keyspace for data replication, migration, and a range of other use cases. CDC drivers are available in Java, Go, and Rust.

 The connector allows you to use Apache Kafka and the Confluent platform while taking advantage of ScyllaDB’s underlying shard-per-core, shared-nothing architecture.

ScyllaDB supports the CQL binary protocol version 3, so any Apache Cassandra/CQL driver that implements the same version works with ScyllaDB, for example, Spark-based applications can use CQL drivers OOTB to work with ScyllaDB.