Was this page helpful?
Advanced Internode (RPC) Compression¶
ScyllaDB Cloud uses internode (RPC) compression to reduce the amount of data transferred between nodes in a cluster.
Internode compression is used for operations such as replication, repair, streaming, and coordination between nodes. These can otherwise generate significant network traffic, especially in multi-node or multi–data center deployments.
Compression substantially reduces the total cost of ownership (TCO) by lowering intrazonal network costs.
Compression Mechanism¶
Internode compression reduces the size of data exchanged between nodes by applying compression before transmission and decompression on receipt. This improves network efficiency and can reduce latency in bandwidth-constrained environments.
ScyllaDB Cloud uses Zstandard (ZSTD)-based compression, a modern algorithm that provides a strong balance between compression ratio and CPU overhead. This allows data to be compressed effectively while maintaining high throughput and low latency.
In addition, compression is enhanced using dictionary-based techniques, which improve efficiency for repeating data patterns commonly found in database workloads.
Behavior in ScyllaDB Cloud¶
ZSTD-based internode compression is:
Enabled for all clusters and service tiers
Applied automatically to internode (RPC) traffic
Fully managed by the platform
No configuration or user action is required.
This behavior ensures efficient use of network resources and contributes to overall cluster performance without requiring manual tuning.
Note
Some dictionary training data may be encrypted using storage-level encryption instead of database-level encryption.