Navigation ↓

Our company

OpenIO was founded in 2015 by a group of engineers, passionate about technology. They had the foresight to develop ConsciousGrid™, an innovative technology that enables OpenIO to scale instantly. And while most object stores fall short when it comes to performance, OpenIO Object Storage was designed from the beginning to provide high performance throughput, making it an ideal choice for Big Data, AI and HPC workloads.

About OpenIO

OpenIO was founded by Laurent Denel (CEO), Jean-François Smigielski (CTO) and four other co-founders, to offer an innovative software-defined object store that builds hyper-scalable, on-premises infrastructures. We envision a data-centric world where OpenIO is recognized as the universal storage solution for unstructured data.

Object storage has always been very focused on low $/GB and high scalability, but some of its characteristics can make it less flexible and performant. From the very beginning, the OpenIO team wanted to change this paradigm, and aimed to create a solution that blends efficiency, flexibility and hyper-scalability with consistent high performance. Our mission today is to deliver an open source, high performance object storage solution that meets the demanding needs of customers working on Big Data, AI and HPC.

OpenIO Object Storage is based on a unique grid of nodes design and ConsciousGrid™ technology that:

  • has dynamic behaviors
  • is hardware-agnostic and supports heterogenous hardware within the same cluster
  • enables instant and unlimited scaling without mandatory data rebalancing;
  • has consistent high performance even during scaling operations
  • has been optimized for Big Data, HPC and AI

We aim to uphold all the promises of object storage, while including the additional benefits that can be found on other platforms.

Proud member of
  • Cloud28
  • EuroCloud
  • M3AAWG
  • iRODS
  • Teratec
  • Impact USA
  • La French Tech

Leadership

Milestones

Timeline:
  • 2006 Idea & initial concept

    In the data center of the company we were working for, data needs were growing exponentially. Our traditional NAS/SAN storage solution was overwhelming the team charged with running these platforms. Simply adding more of the same was not an option. Since most of the data we were dealing with was made up of immutable objects, it was possible to switch to a web-like GET/PUT/DELETE paradigm instead of a POSIX filesystem API. This approach would also make it possible to eliminate the complex hardware required to run these filesystems in a fully distributed manner. This led to the idea of an object storage solution.

  • 2007 Design & development

    We designed a custom solution with a massively distributed directory, which enabled a true scale-out design. Each node would contain a part of the directory, so that the I/O intensive metadata load would be distributed across the cluster. Each new node would be immediately used, so there was no need to re-dispatch existing data to benefit from new hardware. Data storage would be decoupled from metadata storage, providing true scalability and unmatched performance.

  • 2009 First massive production

    In February 2009, we launched a new platform for a massive production of a large-scale email system. We realized that moving data to process it outside the storage platform was non-optimal. The platform had an excess of CPU and RAM, and we decided to run data processing on the object storage platform itself.

  • 2012 Open sourced

    We made the solution an open source project. Originally designed as a general-purpose object storage solution, it has since been used across multiple applications: consumer class email and cloud storage, archiving of radar speed control camera recordings, healthcare data, and much more.

  • 2014 10+ PB managed

    The object storage solution was used to store more than 10 Petabytes of email for two major French Telco/ISPs. These clusters stored 10 Billion objects, with 20Gbps of bandwidth capacity at peak hours. We provided 24/7 support for these critical systems.

  • 2015 OpenIO is launched

    We started OpenIO as a company, to offer the solution as a fully supported product.

  • 2017 OpenIO Raises $5 million

    We raised $5 million from Elaia and Partech.

  • 2018 25 customers, 30 PB in production

    After the first year of general availability, we have 25 customers using OpenIO to store more than 30 PB of primary data in production.

  • 2019 OpenIO exceeds 1Tbps write speed

    We confirmed our position as the best object store for Big Data, AI and HPC when we set a world record of 1.372 Tbps, in a production environment on a Criteo platform.

Investors

  • Elaia
  • Partech
  • OktoCampus
  • Nord France Amorçage

Our partners

  • Atempo
  • Rubrik
  • Backblaze
  • Dracoon
  • SME
  • Vade Secure
  • BlueMind
  • Synacor
  • Huawei
  • Mellanox Technologies
  • Marvell
  • OVH So You Start
  • Supermicro
  • Western Digital
  • 2CRSI
More partners >

Contacts

Headquarter

  • Lille
    2 Bis Avenue Antoine Pinay
    59510 Hem, France
    > Show on map

Japan

  • Tokyo
    WeWork
    Ark Hills South Tower 16F
    1-4-5 Roppongi, Minato-ku
    Tokyo 106-0032
    > Show on map
More ways to reach us
We look forward to assist you. Please tell us how to reach you, and we’ll contact you shortly!

In the spotlight

  • OpenIO: work organization and continuity of service during the COVID-19 epidemic - Preview
    Blog post OpenIO: work organization and continuity of service during the COVID-19 epidemic
    View content
  • Kayrros uses OpenIO for Big Data Storage - Preview
    Case Study Kayrros uses OpenIO for Big Data Storage
    View content
  • Why throughput matters in object storage and how we exceeded 1 Tbps on commodity servers. - Preview
    Blog post Why throughput matters in object storage and how we exceeded 1 Tbps on commodity servers.
    View content