Each Map also introduces its own segmentions, limitations and categorizations.

​

This is the case for the File Storage Map with 3 sub-segments: Enterprise, High Performance and Cloud File Storage. It means that matrix on this topic has an extra tab to cover these file storage categories.

​

Globally, we avoid all offerings with a single file server without any capabilities to cluster and build a scale-up/out or distributed mode. Beyond these points, we consider shared-nothing or shared-everything models as well as disaggregated architectures.

​

Again, we also decided to add strict criteria considering only players who actively promote file storage they design, build and develop. Distributing an entire file storage from an external source prevents this product from participating in the study.

​

This market report lists players offering on-premises file storage solutions both in terms of software or hardware with cloud integration or flavor as an option.

​

We study file storage that powered production data tier considered as primary (file) storage and we don't cover at all copy data tier as secondary storage even if some of the market offerings offer interesting distributed file storage, they’re out of the scope of this report. It also means that some of these products could be used as secondary storage. Again it’s about a role and not a product or technology.

​

For the high performance category, we pay attention if a parallel file system is used or just industry standard file sharing protocols such as NFS or SMB. Lustre integration or other open source elements are also scrutinized. Same remark for Posix compliance and dedicated data services needed in the domain. Cluster size, storage capacity, node counts… are also considered. Workloads, use cases are critical here with validation of specific applications known in the domain. Presence in the IO500 is important as well.

​

For the enterprise category, we consider industry standard file sharing such as NFS and SMB. Enterprises’ use cases are the ones considered here. Maturity of the solution is important and external solution integration like oeming some file sharing component is checked by the analyst team.

​

It’s fundamental to understand the differences between an external and internal file system like a parallel file system or NFS to illustrate the notion.

​

For the cloud category, back-end support with private or public cloud services with S3 or other APIs are considered. It could be public players such AWS, Azure or Google but also several others. The access methods exposed as NFS and SMB are a must in this category. Multi-site, multi cloud providers and globalization or unification of file systems are also considered.

​

For every category, we insist on the data protection and data services features offered by players active in that domain. They are different in each category.

​

Unstructured data also is associated with a S3 API so this element is considered, same as access to the same content with file- or object-based access methods. HDFS offerings and players are excluded from this report.