Analyst firm Gartner released its latest version of its Magic Quadrant for object storage that attempts to rank and organize the distributed file system and object storage system market. Instead, it really creates more questions and confusion instead of providing any clear direction. Unless, of course, you believe a rapidly emerging market can be better served by legacy vendors than it can by new vendors.
The foundational problem with the Gartner Magic Quadrant is that it covers too broad of a spectrum of systems. Ironically, it is just the opposite of our complaint about the way Gartner ranks primary storage by separating out all-flash systems. Instead, in secondary storage, which is actually more diverse, Gartner tries to do too much by combining object storage and distributed file systems into a single grid.
While all object storage systems have distributed file systems that allow them to aggregate multiple nodes into a single storage pool, not all distributed file systems have to use object storage. This means the use cases are for too diverse to include in a single grid. There are high performance distributed file systems and there are file systems that are more for the archive use case.
Breaking down the Quadrants
The “leaders” quadrant is the least occupied of any Gartner Magic Quadrant we’ve seen. Two of its three members; Dell/EMC and IBM have multiple (at least three each) systems/solutions that can qualify for consideration. For example, the report specifically calls out Isilon as a dominant market leader, with 50% of the overall revenue in the space. While Isilon is a solid product, its use cases range the gamut from primary storage, to archive storage to data protection storage. Counting all Isilon revenue in the grid and using it to justify EMC’s positioning seems unfair.
The “visionary” quadrant is also sparsely populated and again full of companies with such a mix of use cases that is makes it hard to understand. For example, two (SwiftStack and Qumulo) of the three companies in the visionary space couldn’t have more opposite use cases. SwiftStack is typically used in native object storage use cases like long term data retention and archive. Qumulo is much more focused on high performance use cases. Red Hat, similar to IBM and Dell/EMC, have multiple solutions ranging from primary storage to long term data retention.
The “challenger” space is the least populated of the four quadrants with one participant, Hitachi Vantara. If IBM and Dell/EMC are in the leadership space, it’s confusing how Hitachi can’t be in that space. They are a large company and Hitachi Vantara is being very successful with the product, especially recently. A fact that Gartner’s own report points out.
It is also confusing how NetApp can be in the “niche” quadrant and not paired with wherever Hitachi Vantara is. They are in a similar situation as Hitachi, both use a single object storage system to cover the market. The only real difference is that it appears Hitachi is being more successful with its object storage solution than NetApp. Again, calling into question how Hitachi received a lower ranking than NetApp.
The “niche” section is the most crowded section of the quadrant. This makes sense considering the whole market itself is somewhat niche. Most organizations are still trying to figure how exactly they should use object storage and distributed file systems. And, again, they are using object storage and distribute file systems for an incredibly wide range of use cases.
Once again the positioning of the companies in this space is suspect. For example, Caringo, Cloudian and StorageCraft (Exablox) have all been very innovative in their respective spaces. All three deserve to be in the visionary quadrant. SUSE and Red Hat use very similar core code, how they can be in almost opposite quadrants is confusing.
Missing in Action
Equally confusing is the companies missing from the quadrant all together. Part of this omission may be because of Gartner’s requirements for making the list, 50 production customers each consuming more than 300TB’s or $10 million in revenue in the space. But it seems that Gartner should include companies like Rubrik, Hedvig and Cohesity. None of these companies market themselves as object storage systems, but they are all clearly distributed file systems. We also believe that each has more than 50 customers although the capacity requirement may be tripping them up.
Scale-Up is Out?
Also confusing to us is the automatic elimination of a scale-up solution especially on a list that essentially includes everything else. There are many scale-up storage systems that can meet the 300TB capacity requirement and many customers that don’t need much more.
Fixing the Quadrant
The problem with the Gartner Quadrants, and this one especially, is it is focused on the technology and not the use case. Customers are not typically shopping specifically for an Object Storage System or a Distributed File System, instead they are trying to SOLVE A PROBLEM. Why not build quadrants by use case instead of by technology? For example, why not have an Archive Quadrant, a Backup Hardware Quadrant or a High Performance File System Quadrant? That’s how IT professionals make decisions.