The terms SDS (Software-defined Storage) and SRM (Storage Resource Management) are often used interchangeably, but in fact are wholly different. One (SDS) is a fairly new term, while the other (SRM) has been around for decades. In this blog, we’ll discuss the distinctions between each term.
SDS is a fairly new term, and as such is often misused and misunderstood. Luckily, most definitions of SDS always seem to include the notion of a control plane and a data plane, so we’ll adopt that idea as well. The concept of “planes” came from the networking industry (specifically from Nicira) circa 2007 in the context of Software-defined Networking (SDN) and this idea was adopted by the storage industry a few years later.
The main idea behind both SDN and SDS is that inexpensive physical devices should be tasked only with the simplest details of moving and storing data (the data plane), whereas a higher-level intelligence software layer would instruct these devices where and when to move data (the control plane). For SDS, this uncoupling of the storage hardware and software led to many designs built around commodity storage devices and sophisticated control software, some examples being VMware’s EVO:RAIL and Red Hat’s Ceph.
SDS is an important concept for companies that design storage systems and NetApp has been one of the early adopters of this approach. Cloud ONTAP and Data ONTAP Edge are two examples of NetApp products that utilize the SDS model by running the Data ONTAP storage operating system as a virtual control plane on top of commodity hardware; in the former, hyperscale clouds are the commodity platform, while the latter is based on industry-standard x86 hardware. It is important to point out, however, that SDS is typically limited to controlling storage devices within a single architectural domain – in other words, vendor A cannot necessarily control storage products made by vendor B. That’s where SRM comes into play, as we’ll see in the next section.
Storage Resource Management
SRM has been around since the days of the mainframe (see this 1991 Computerworld article), and should not be confused with SDS, as storage “control” is much different from storage “management”.
Unlike SDS, SRM involves techniques that help optimize the overall utilization of all available storage capacity. Some basic functions of SRM software include storage performance monitoring, capacity planning, policy creation and enforcement, and chargeback usage reporting.
As shown in the diagram below, SRM can be thought of as a 3rd plane, dedicated to management and functioning above the SDS data and control planes. A simple analogy from the airline industry can be used to describe the relationship between the three “planes” (no pun intended): the data plane can be thought of as the physical aircraft that carries fuel and passengers, while the flight crew can be thought of as a control plane that safely transports the passengers from point to point. Observing all the aircraft in flight is an air traffic control infrastructure, which includes of a series of control towers that monitor and optimize all activity within their domain – in other words, a management plane.
Modern enterprise storage infrastructures can utilize these same three planes. As shown on the right of the diagram, in NetApp’s case, the data plane consists of data storage capacity provided to enterprise customers consisting of either hardened data center class equipment, commodity off-the-shelf hardware, or cloud based storage services.
NetApp’s leading storage operating system, Clustered Data ONTAP, is the basis of its control plane – utilizing a data fabric that connects various public and private cloud storage end points into a coherent, integrated, and compatible system—in essence, a single virtual storage cloud with a unified data structure.
SRM - A Control Tower for Enterprise Data
Just as an airport control tower monitors all aircraft within its vicinity, NetApp OnCommand Insight monitors all data within an enterprise IT organization, and thus resides on the management plane. OnCommand Insight collects an enormous amount of information that is stored within its data warehouse, leveraging a high-performance Cassandra database, where advanced analytics are used to perform troubleshooting, performance analysis, capacity, and change management.
By being able to collect and monitor all the real-time performance metrics that are being generated within a heterogeneous IT infrastructure, organizations can quickly identify potential problems at a server, networking, or storage level and make the right business decisions faster to remediate those problems.
In upcoming blogs, we’ll take a deeper dive into the capabilities of OnCommand Insight and compare its features to other SRMs on the market today. In the meantime, you can whet your appetite by watching the videos below that highlight the various capabilities of OnCommand Insight.