OpenIO Object Storage
-
Hyper-scalable object storage
Scale seamlessly from Terabytes to Exabytes. Simply add nodes to expand capacity, without rebalancing data, and watch as performance increases linearly.
-
Designed for high performance
Transfer data at 1 Tbps and beyond. Experience consistent high performance, even during scaling operations. Ideal for capacity-intensive and challenging workloads.
-
Hardware agnostic, lightweight solution
Use servers and storage media that suit your evolving needs. Avoid vendor lock-in. You can combine heterogenous hardware at any time, of different specs, generations, and capacities.
- Flexible and instant scaling
- Optimal positioning of data
- Performance optimization
- Smart hardware management
- Fine-grained storage policies for automated tiering
- Object versioning and container snapshots
- On the fly and delayed compression
- WORM mode
- Erasure coding and n-time replication with multiple dynamic storage policies available
- Geo-distributed and stretched cluster topology
- Encryption of data at rest
- S3 and Swift APIs
- Native Python, C and Java APIs
- File access through OIO-FS
- Big Data integrations
-
Kayrros
-
20Servers
-
2 PBStorage
-
1 GBAvg file size
-
40%Savings
We recreated the public cloud provider offer, using our own services and by building our own private cloud, not accessible from the public network. In doing so, we have improved our performance, and we have reduced our costs by 40%.
Read the Customer Story -
-
Dailymotion
-
50 Mvideos per month
-
800 Mobjects
-
40 PBstorage
-
30%storage growth per year
OpenIO stands out as a scalable, efficient and easy-to-run object store. It costs us 50% less than a comparable scale-out NAS solution.
Read the Customer Story -
-
CEA
-
4UARM
-
96nano-nodes
-
768 TBnet storage
-
S3
OpenIO perfectly meet the storage needs of an HPC platform. We can process and store hundreds of billions of files of all sizes, and OpenIO maintains a linear high performance, even when we add new capacity.
-
-
IIJ
-
14 Memails
-
5,5 Kemails indexed per minute
-
2sites
-
100%flash
With peak requests that can easily reach 5,500 emails/sec, we needed a high performance object store. With OpenIO we were able to build a very high performance all-flash storage system.
Read the Customer Story -
Highlights
-
Stateless native object store
OpenIO stores objects within containers in a persistent 3-level massively distributed directory: there is no classic subdirectory tree, objects are stored in a flat structure.
The data query path is independent of the number of nodes, and performance is not affected by the growth of the storage capacity.
-
Grid of nodes architecture
The OpenIO cluster layout is organized as a massively distributed grid of nodes.
This creates a flexible, resourceful solution. OpenIO doesn't require data rebalancing when you change the topology of a cluster, and it permits efficient load balancing on heterogeneous nodes.
-
ConsciousGrid™ technology
ConsciousGrid™ is an innovative self-learning technology that uses real-time metrics from the nodes (CPU, I/O, capacity...) to compute a quality score and handle any requests accordigly, matching it with any constraint (e.g. respect of storage policies).
-
GridForApps™
OpenIO catches all events that occur in the cluster and can pass them up in the stack.
OpenIO can also perform real-time event-based metadata indexing and search directly on the storage infrastructure, more functions can be activated on customer demand.
-
Open source core
The core of OpenIO Object Storage and the object APIs are open source and hosted on GitHub.
This guarantees the long-term availability of the solution and gives your engineers the opportunity to understand how our technology works.
Want to get started?
Here’s what you can do with OpenIO: build a hyper-scalable, on-prem storage infrastructure that is flexible and fast, and at a fraction of the cost of a public cloud solution.
Talk with us