Xatastor: Storage engine
for very high multitenancy
We have developed an innovative storage engine optimized for agentic workloads. It is ideal for running a very large number of PostgreSQL instances with optimal resource usage.
Data branching
Built-in Copy-on-Write storage powering instant branches and snapshots.
Separation of storage and compute
Xatastor volumes are exposed over the network via the NVMe-oF (NVMe-over-Fabrics) protocol.
Bottomless storage
Thin provisioning means that customers are billed strictly by the storage space they actually use.
High scalability and multi-tenancy
Xatastor is optimized for a very large number of lightly used volumes. At the same time, it supports high performance on each of the individual volumes.
It scales to tens of thousands of volumes and snapshots on each storage node. This means it scales to millions of volumes with minimal hardware requirements.
- Inactive volumes only consume disk space
- Over 100,000 volumes and snapshots per storage node
- Scalable to millions of volumes and snapshots in total
Built on industry-standard technologies
NVMe is a high-speed storage protocol that maximizes storage performance through low latency and parallel processing. NVMe over Fabrics, including NVMe/TCP, extends these benefits over network connections, enabling scalable and efficient remote storage with minimal performance loss.
- High-performance parallel processing
- Continuous checksum verification and recovery
- Reduced context switching and CPU overhead
- Zero-copy data transfer
ZFS is a combined file system and logical volume manager designed from the ground up around data integrity. Its copy-on-write transactional model keeps the on-disk state consistent, while end-to-end checksumming detects silent corruption and, with redundancy, repairs it automatically. Among its standout capabilities are zvols, which expose a dataset as a raw block device backed by the pool.
Our user-space NVMe-oF implementation minimizes context switching and CPU overhead, making it ideal for high-performance storage applications such as databases.
Separated storage and compute architecture
The Xata platform combines well-known and tested technologies with an innovative storage system to offer a high-performance, scalable, and secure PostgreSQL service.
Flexible auto-scaling compute and storage nodes
The PostgreSQL instances run inside a Kubernetes cluster. Because storage is separated, auto-scaling is free to continuously optimize the infrastructure underneath.
Logical storage volumes mounted via CSI
A Kubernetes CSI driver is used to mount the logical storage volumes to the PostgreSQL pods.
NVMe over Fabrics (NVMe/TCP)
The storage volumes are connected to the PostgreSQL pods via NVMe-oF, which is a high-performance network protocol for block storage.
Distributed storage cluster
The storage cluster uses multiple storage nodes (e.g. EC2 instances with either local NVMe drives or EBS drives), and the volumes are automatically distributed across them.
Postgres for agentic workloads.
Use your existing Postgres. Run it better with Xata.