Scylla Open Source

Get better, more consistent performance and lower costs while maintaining the high availability traits and scale-out database designs of Apache Cassandra® and Amazon DynamoDB®.

Advantages of Scylla Open Source NoSQL

Teams around the globe trust Scylla at the heart of their big data architectures.

icon-latency-1

Consistent, Low Latencies

Scylla’s underlying architecture enables it to consistently deliver P99 latencies in the low single-digit milliseconds, even under extreme load. Maintenance operations do not slow performance.

icon-throughput-1

10x Higher Throughput

Scylla uses a highly asynchronous, non-blocking design that takes full advantage of modern hardware architecture. It provides the throughput you need for even the most taxing workloads.

Less Complexity

Scylla has a unified cache and self-optimizes its performance, making it easier to use than Cassandra, which requires complex manual cache and JVM tuning.

icon-lowcost

Low Total Cost of Ownership

Scylla makes full use of your hardware infrastructure, minimizing node count and reducing administrative overhead.

Database Monsters of the World Connect!

VIRTUAL CONFERENCE | JAN 12-14

Designed for Real-Time Big Data Use Cases

An open source NoSQL database that's purpose-built for modern applications.

icon-shard-per-core-1

Shard-per-Core

Scylla’s shard-per-core architecture equally divides an instance’s resources, providing an equal slice of CPU, RAM, and storage to a slice of the database.

icon-close-to-metal-1

Close-to-the-Metal

Built on top of Seastar, an engine written in C++ that leverages low-level Linux capabilities for async communications, memory management, scheduling, prioritization and caching.

icon-numa-2

NUMA Optimized

Scylla’s assembly machine code ensures the most-efficient use of modern multi-core, multi-CPU NUMA hardware. This allows Scylla to scale vertically and fully utilize the largest servers found in public clouds.

icon-async-1

All Things Async

Scylla uses an “all things async” architecture, using C++ futures and promises to ensure each sharded process of Scylla can operate efficiently and independently.

icon-unified-cache-1

Unified Cache

Scylla’s unified data cache means the most frequently accessed data will be available immediately in memory. In comparison, Cassandra uses multiple cache layers which adds complexity and requires manual tuning.

icon-scheduler-1

I/O Scheduler

Even using fast NVMe disks you don’t want your database bounded by disk I/O. Scylla provides an advanced I/O scheduler to manage concurrency, balancing overall throughput with lowest latencies while accessing persistent storage.

icon-self-optimizing-1

Self-Optimizing

Scylla was designed with many self-tuning and automatically adaptive capabilities to ensure ease-of-operations. There’s no JVM to tune. This reduces administration from initial set-up and through peak operations.

icon-c-ecosystem-1

Full C* Ecosystem

Scylla is fully compatible with the full set of drivers and connectors for Apache Cassandra, whether you program in C, Java or Python, and whether you need to connect to Apache Spark or Apache Kafka.

Supported Interfaces

Scylla supports two different interfaces: A Cassandra Query Language (CQL) interface and an Amazon DynamoDB-compatible API. We recommend our CQL interface, which supports our full feature set, for all users aside from those who need to maintain DynamoDB API compatibility.

Scylla is a drop-in replacement for Cassandra. Keep all your existing client code — you can use your existing data models and queries using our CQL interface for Cassandra-like operations. Aside simply performing far better, Scylla’s implementation of CQL goes beyond Cassandra, adding new and improved features, like:
You get all this while keeping your development ecosystem the same: client drivers and big data frameworks like Apache Kafka® and Apache Spark™ all connect to Scylla seamlessly. Scylla’s Cassandra-compatible CQL interface provides three key benefits to users:
  1. Cost Scylla can run your workloads more efficiently, meaning you do get the most out of your infrastructure. Even though both Scylla and Cassandra are free open source NoSQL databases, the hardware you require to run them on is not. Because Scylla can scale better vertically, you can run it on bigger nodes, reducing your node count to save further on related operational expenses. According to our benchmark, with Scylla users can save up to 2.5X on their cloud vendor bills.
  2. Performance Scylla was implemented in modern C++ to support features such as instructions per cycle; lock-free, log structured cache; heat-based load balancing; workload prioritization; userspace schedulers and much more. This results in improvements to latency and throughput, and allows Scylla to scale both vertically and horizontally. Scylla can take advantage of high-density nodes, some of which can now support hundreds of cores per server.
  3. Openness Like Cassandra, Scylla is an open source NoSQL database. Users can review the source code and any known defects, and in return can add their own contributions to the project. Operationally, Scylla can run on any suitable server cluster regardless of location (on-premises, in a private or public cloud, or our own Scylla Cloud DBaaS) or deployment method (bare metal, containerized, virtualized, or via Kubernetes). This contributes to the user’s lower TCO by allowing deployment flexibility in line with their existing operations.

Scylla is a drop-in replacement for DynamoDB. No need to change your client code — you can keep all your existing data models and queries when using our DynamoDB-compatible API. Scylla’s implementation of the DynamoDB API goes beyond Amazon’s implementation with leading features such as:

  • Leaderless active-active architecture avoids single point-of-failure
  • Optional fast writes that avoid costly read-before-writes

All while keeping your development ecosystem the same: SDKs and big data frameworks like Apache Kafka® and Apache Spark™ all connect to Scylla seamlessly.

Scylla’s DynamoDB-compatible API provides three key benefits to users:

  1. Cost DynamoDB charges for read and write transactions (RRUs and WRUs). A free open source NoSQL database solution eliminates these costs and minimizes other operational expenses. Scylla’s design efficiency allows developers to use significantly fewer resources for the same task or workload. According to our benchmark, users can expect to save 80% – 93% overall to support the same workload (5x-14x less expensive). Because of its highly performant design Scylla would also eliminate the cost of any in-memory cache (such as DAX).

  2. Performance Scylla was implemented in modern C++ to support features such as instructions per cycle; lock-free, log structured cache; heat-based load balancing; workload prioritization; userspace schedulers and much more. This results in improvements to latency and throughput, and allows Scylla to scale both vertically and horizontally. Scylla can take advantage of high-density nodes, some of which can now support hundreds of cores per server.
  3. Openness Scylla has an active and growing open source community. Users can review the source code and any known defects, and in return can add their own contributions to the project. Operationally, Scylla can run on any suitable server cluster regardless of location (on-premises, in a private or public cloud, or our own Scylla Cloud DBaaS) or deployment method (bare metal, containerized, virtualized, or via Kubernetes). This contributes to the user’s lower TCO by allowing deployment flexibility in line with their existing operations.

Join the Scylla Community

icon-connect-1
Connect with your fellow users and developers.
icon-best-practices-1
Learn to apply best practices.
icon-contribute-1
Help improve our code or contribute your own