File versus Object: implementing a unified object storage platform for both
Originally presented on February 1, 2017
The growth of unstructured data is no longer news, but it will continue to remain as a challenge to most. Many corporations find themselves dealing with volumes of data that far outstrip anything they have ever dealt with in the past. While cloud and object based storage platforms have advantages when dealing with the massive data volumes, enterprises find themselves with legacy applications, that have been build around the POSIX file system paradigm. Thirty years ago when the standard was born, hard drives were measured in megabytes. Since that time much has changed, but file systems have remained at the very core of how applications store data, and notably unstructured data, where the only structure around the data was the file system.
Scality developed a highly scalable object based platform, back when object based storage was in its infancy, but it quickly became clear that, no matter how much systems architects liked tag object based model, the applications would take decades to adapt the new models. Faced with this reality we pursued the development of a true scale-out file system, a file system built on top of a an infinitely scalable object backend.
Another trend accompanying the growth of the volume of data stored, is the increase in the perceived value of the stored data, as machine learning, data mining, collaborative and web 2.0 services with always on, always available have grown to be the expectation.
- What true “scale-out” looks like, and the difference it makes managing large data sets
- How to implement and manage always-on storage, when always means always?
- Data protection at multi-petabyte scale
- Multiple interfaces for multiple workloads
- Geo-distribution models that keep your data available
- Living in a post backup world
- Preparing for the future, be it file or object
Director Product Management and Customer Experience
Chief Engineering Architect