The Coming Software-Defined Storage Revolution


Want the latest storage insights?

Download the authoritative guide: Enterprise Data Storage 2018: Optimizing Your Storage Infrastructure

Share it on Twitter  
Share it on Facebook  
Share it on Google+
Share it on Linked in  

Software-defined storage (SDS) is set to revolutionize the storage industry over the coming years.

But this shouldn't come as a surprise: after all, software-defined computing—in the form of server virtualization—has radically altered the server landscape, and software-defined networking is already beginning to change the way virtual machines are linked together.

Storage is just the final part of the puzzle to get the "software-defined" treatment. "One of the advantages of virtualization is that it enables you to mix and match resources as required. Software-defined storage is the logical extension of this and thus inevitable," believes Mark Peters, a senior analyst at Enterprise Strategy Group.

What Is SDS?

What exactly is software defined storage?

Right now there's no standard definition, but it involves separating storage features from the storage hardware itself. IDC defines software-based storage as "any storage software stack that can be installed on any commodity resources (x86 hardware, hypervisors, or cloud) and/or off-the-shelf computing hardware, and used to offer a full suite of storage services and federation between the underlying persistent data placement resources to enable data mobility of its tenants between these resources."

So SDS enables companies to buy heterogeneous storage hardware—choosing whatever is best value (in terms of price, reliability or any other criterion) at any given time—and then buy and use the storage software that provides the features they need. "SDS means separating the hardware buying decision from the software buying decision," Peters explains.

Likely SDS Vendors

There is no shortage of hardware vendors, but the obvious question is where will the storage software come from? Laura DuBois, a storage expert at IDC, predicts it will come from three key camps:

  1. Open source development—this includes Hadoop with HDFS, OpenZFS, Openstack Swift and storage-specific solutions like Ceph.
  2. Storage ISVs - Companies like Inktank, Nexenta, Red Hat or Symantec
  3. Existing storage vendors that have traditionally delivered storage systems and appliances. "Basically, any storage software stack that could run as independent software on industry standard hardware," says DuBois.

If you look at those three camps, you'll see that they are very similar to the groups that have appeared to provide software-defined networking solutions. There are open source projects like Nox and Floodlight that have produced software-defined networking controllers, and there are commercial vendors like Big Switch Networks and Nicira (now part of VMware.) But perhaps most significantly, the big networking hardware vendors like Cisco and Juniper are all getting involved in the software-defined networking space as well. They have to—if they don't, they may become irrelevant.

What's interesting about the hardware vendors like Cisco and Juniper is that although they are embracing OpenFlow, a software defined networking standard, they are also "propriatizing" their offerings so that their software actually works better if it happens to be used with their own networking hardware.

Something similar is likely to happen in the software-defined storage space, Peters believes. "Hardware vendors will inevitably try and propriatize the technology," he says. "You will still get the 'better together' argument: 'we offer SDS, but integrated together with a system.'"

You can't really blame these storage companies for attempting to preserve their markets, but it's not clear that that route is the one that leads to the biggest benefits to storage purchasers. It's more likely that open source projects or storage ISVs will provide these, says Ashish Nadkarni, a research director in IDC's Storage Systems research practice.

Submit a Comment


People are discussing this article with 0 comment(s)