The New School of Storage: Scality Rides the Unstructured Data Wave


The new school of storage is like any new artistic genre or intellectual movement. It grows out of organic disruption. It’s the offshoot of chaos, having its own impact on our culture and way of life. For our purposes, data is the disruptor. The chaos is in the technology stack and the innovators are the upstarts of the new school of storage.

Disruption tends to surface the ugly side of who we are. The older, more established players will often hide their illnesses and appeal to the power barons.  The young turn to the community. The trick is to build a community who know the old way may work but would prefer to adopt something new that is relevant to today’s workloads. Get enough momentum and the strategy can prove disruptive in itself.  Succeed and you will most likely get acquired. If lucky, your technology will become a standard for the market.

The Scality Approach

Scality is a two-year-old startup that represents this new school of storage.  They appeal to the early adopters, the ones who want to learn about the new stuff and are willing to buy it, too.

The trick for Scality? Patience and persistence. It will take years to mature before the most advanced storage systems see wide scale adoption.

Scality has to take this approach. Just look at its latest news. It’s exciting and new. And that’s the problem. Most IT buyers don’t want to to buy the exciting and new storage technology.

Scality’s technology cuts to the heart of the issue that companies face with unstructured data.  As Scality points out in its news, unstructured data now represents the bulk of storage required by service providers and large enterprises. At the current pace, unstructured data will soon reach 80% of all storage requirements.

Scality’s organic ring storage creates a “distributed, dynamic table that scales by placing the data in containers that are not constrained by the limitations of file-based storage systems.” The system, according to the company, scales and adapts to different storage requirements. Scality executives see this as similar to the way organic systems adapt and replenish themselves. It’s designed to fail, a hallmark of Web scale systems. With Scality, the applications continue to function when servers fail. The load is distributed elsewhere to different parts of the ring storage network.

With RING Organic Storage 4.0, Scality is furthering its claim that distributed storage will replace network attached storage (NAS) as the need for scale continues to grow.

Scality CEO Jérôme Lecat says it can handle hot data, deliver a cost of ownership comparable with Web infrastructures and a management system far simpler than NAS.

Enterprise Strategy Group had this to say :

Until now, object storage could deliver the scalability, but not the performance. Standard response times for object-based storage are in the hundreds of milliseconds to full seconds, much too slow for today’s cloud-based services. In comparison, ESG testing demonstrated that Scality object storage on Intel Xeon servers  equipped with Intel Enterprise SSDs and low-latency 10GbE network environment delivers 4-10 millisecond performance—10 times faster than other scale-out systems. The Scality solution delivers scalable object storage in which performance is never a barrier. Users can solve any performance problem with this architecture just by selecting the right number and types of server nodes and disk.

The demands that data places on an infrastructure means companies of all varieties are re-examining their storage needs and looking at new technologies like those from Scality.

Wikibon Co-Founder David Floyer made the point in an interview yesterday that persistent storage will be the most pressing issue over the next four to five years for IT buyers.  The enti

re stack is changing. Apple, Microsoft and other massive device makers are forcing shifts. x86 processors or ARM – which will be most important? Apps are at the center of the battle with their proliferation and huge appetite for data.

All of this change means a huge opportunity for companies like Scality and the rest, including CaringoClever Safe and fresh entrants such as Basho, which recently introduced  Riak CS, a new software platform for building multitenant cloud storage systems. The big guns like EMC Atmos Nirvanix, Dell, Amazon Web Services S3 and Rackspace Cloud Files  also play in this market as does Gluster, the open-source startup acquired last year by Red Hat, now known as Red Hat Storage.

These companies all have an opportunity, considering the massive influx of big data and the impact it has on how companies keep their data, share it, integrate into apps to do analysis or otherwise.

“The change in the stack is profound,” Floyer said.

Scality is designed for consumer services. Time Warner Cable is building out on Scality to manage its email infrastructure. Email storage is not rocket science. Most of it is archiving. But it is also one of the most demanding of applications due to its unstructured nature. An email can be very small, a few kilobytes in size or very big as in a 20 megabyte attachment. People expect real-time results that can’t be had with a SAN as effectively as they can get with distributed storage that horizontally scales. TWC has more than one petabyte of data. It is growing 40% a year. For TWC, it just became economically impractical to store email on high-end SAN machines.

Lecat says the company wins these kinds of deals by connecting with communities that understands the innovations that come with distributed storage. And these are early days in the market.  In Lecat’s view, the shift will take ten years. It is just the beginning of that shift so it makes no sense to waste time with people conditioned over the past 20 years to believe that storage products from EMC or NetApp are the only answer.

But many enterprise and storage architects will go with the technology they know so well. It may not make sense to invest in legacy storage technology. But often these customers have long established relationships with well-known providers. And these providers have massive marketing budgets to convince their customers that they need to invest in the cloud and other buzz words.

“Our technology is very different than what EMC or or VCE does.” Lecat said. “What we do from a market standpoint – we find these people that realize that EMC is not giving them the entire picture.”‘

An EMC spokesperson declined to comment for this story.

Still Not Cool Enough?

But is Scality next generation enough? Dreamhost is spinning out its storage technology into a separate company. The storage technology is based on Ceph, a new distributed file system developed by Dreamhost Founder Sage Weil. Ceph is getting considerable attention and is being discussed as an alternative replacement for OpenStack Swift storage.

The 451 Group published a short report about Ceph late last month. It cited Ceph’s scale-out architecture and its ability to provide file, block and object storage. It is expected to attract interest from service providers and enterprise providers looking to build out cloud storage services. The Ceph software has attracted a fairly large following for its flexibility and scalability.

Companies like Parallels Software are looking at Ceph as an alternative to Scality, which is perceived by some as a first generation technology in the vein of Amazon S3.

“Scality is a first generation of public storage,” said James Bottomley, CTO of Server Virtualization for Parallels. “It realistically looks like s3 storage. The main problem we see is it does not support sufficient updating of the objects of the image in a virtualized environment.”

But when pressed, Bottomley said this is something that no one company can claim has a clear advantage. Scality may lack that capability but so does most of the competition.

As 451 points out, the growing success of Amazon’s S3 and EBS cloud storage services is putting pressure on service providers and enterprises to build out their own offerings.

But the greater challenge? Having the patience and persistence to drive its presence in the market by connecting with early adopters.

Challenges for the The New School

The members of the new school of storage face a challenge. And that’s convincing the market that data in itself has a value of one.

Shared data? As Floyer points out, the value of data scales the more you share. Your data sets go far further when aggregated with data sets from other organizations. It’s far richer in scope.

“You need apps that can manage vast amounts of data,” Floyer said. “Object storage becomes central to that.”