DAOS

Spec sheet

Purpose

DAOS is an open-source, software-defined storage system for AI and HPC workloads.Read 3 Oct 2026 · daos.io

Audience

The site presents DAOS as an open-source storage platform for AI and HPC workloads.Read 3 Oct 2026 · daos.io

Integrations

The DAOS client ecosystem includes libdfs for POSIX namespaces, and its user guide documents a PyTorch interface for loading data from DAOS.Read 2 Oct 2026 · docs.daos.io

Security

DAOS supports access control lists on pools and individual containers.Read 2 Oct 2026 · daos.io

Support

The DAOS site lists member support by email at [email protected] and links to a Linux Foundation support page.Read 3 Oct 2026 · daos.io

License

The project repository displays the BSD-2-Clause Plus Patent License.Read 3 Oct 2026 · github.com

Access control

Fine-grained ACLs can be set on storage pools and individual containers.Read 3 Oct 2026 · daos.io

Data model

Its core interface is a scalable key-value-array store, with data models including POSIX namespaces and HDF5 datasets built over it.Read 2 Oct 2026 · daos.io

Data models

The overview lists POSIX namespaces, HDF5 datasets, and Python among data models built over libdaos.Read 3 Oct 2026 · daos.io

Data protection

Objects can use erasure coding, replication, or no data protection; DAOS can trigger self-healing rebuilds when an engine becomes unresponsive.Read 2 Oct 2026 · daos.io

Datasets and snapshots

DAOS containers represent datasets, can be snapshotted, and support fine-grained access control.Read 2 Oct 2026 · daos.io

Deployments

DAOS can be deployed on premises or in cloud environments.Read 3 Oct 2026 · docs.daos.io

Framework support

The DAOS v2.8 documentation lists support topics for MPI-IO, HDF5, Python, PyTorch, Spark and Hadoop, and TensorFlow-IO.Read 3 Oct 2026 · docs.daos.io

Network support

DAOS can use libfabric or UCX and lists Ethernet, RoCE, InfiniBand, Omni-Path, and Slingshot fabrics.Read 3 Oct 2026 · daos.io

Notable limits

DAOS 2.8 does not support Ubuntu, Fedora, CentOS Stream, or RHEL 8 and its clones.Read 2 Oct 2026 · docs.daos.io

Performance

The site describes distributed metadata, ultra-high bandwidth and IOPS, and time-to-first-byte below 20μs when using RDMA where available.Read 3 Oct 2026 · daos.io

Protection and recovery

Objects can use erasure coding or replication, and DAOS can trigger self-healing rebuilds when an engine becomes unresponsive.Read 3 Oct 2026 · daos.io

Release

The DAOS documentation identifies its current documentation set as version 2.8.Read 3 Oct 2026 · docs.daos.io

Snapshots and scale

DAOS containers are user-managed datasets that can be snapshotted; the site says a POSIX container can hold trillions of files or directories and petabytes of data.Read 3 Oct 2026 · daos.io

Storage interface

Its baseline interface is a high-performance, scalable key-value-array store exposed through libdaos.Read 3 Oct 2026 · daos.io

Storage pools

Pools can span all or some storage engines, and pool access can be controlled with fine-grained ACLs.Read 2 Oct 2026 · daos.io

Supported platform

DAOS 2.8 is built for Linux on x86_64; its supported server and client distributions include SLES 15 SP6/SP7, openSUSE Leap 15.6, and EL 9.6/9.7.Read 2 Oct 2026 · docs.daos.io

Where Specifiction places it

• Software-Defined Storage Software#16 of 307.7
• Storage Replication Software#13 of 247.6