SDS - Defining a Term that Eludes Definition

by Jaap van Duijvenbode on March 12, 2017

What is Software Defined Storage (SDS)? That question is becoming a key issue as the pressure to change the way data has traditionally been stored and delivered to users grows more intense. The amount of information companies have to deal with continues to grow at exponential rates, and traditional storage solutions are becoming more and more inadequate to the challenge. The software defined paradigm is being hailed as the answer to the problem.

In a sense, industry adoption of SDS seems to be running on two separate tracks at the same time. On the one hand, there’s a lot of interest, and actual product purchases are growing at a high rate. According to market research firm MarketsandMarkets, the market for SDS products is growing at a Compound Annual Growth Rate (CAGR) of 36.7%, and can be expected to reach $22.56 billion by 2021. The other side of the coin, however, is that there’s still a lot of confusion among IT professionals about the SDS value proposition. In a survey conducted by ManageEngine, 47 percent of respondents indicated they were unsure of whether SDS would be an asset to their storage operations.

Is SDS Just A Marketing Buzzword?

Some storage professionals still question whether SDS is even real. They take note that the Wikipedia article on SDS begins by referring to it as a “marketing term.” Even the Storage Networking Industry Association (SNIA) has called SDS “a marketing buzzword.” Yet, such references to marketing are not intended to dismiss the SDS concept as nothing more than promotional hype. Instead, they serve to highlight the confusion that still exists among many in the storage industry about just what the term “software defined storage” really signifies.

The main source of that confusion is the fact that, as often happens with hot new technologies, some vendors have raced to hop on the SDS bandwagon by simply relabeling existing storage products as being software defined. The natural result is that each vendor pushes a definition of SDS that somehow seems to line up perfectly with the features of its own offerings.

A Working Definition For SDS

Although there is as yet no “official” definition of SDS, there is widespread agreement about the underlying concepts embodied in the term. The fundamental feature of the software-defined paradigm is the virtualization of storage. Intelligence is shifted from hardware into an overarching layer of software that manages the entire storage infrastructure as a single entity. The control and data access mechanisms of the storage system are decoupled from the underlying hardware and presented to users through a unified software interface that is consistent regardless of the nature and even the geographical locations of the actual physical storage units.

The SNIA definition of SDS sums it all up by declaring that SDS is “virtualized storage with a service management interface.”

Fundamental Features of SDS

In its definition, SNIA identifies several features that it considers necessary elements of any legitimate SDS implementation. These include automation, standard interfaces, virtualized data paths, seamless scalability, and transparency. Let’s take a brief look at each of these.

Automation: Because SDS treats all the devices it manages as a single pool of storage, regardless of the types of devices or media that may be included, critical storage system tasks can be automated and applied system wide. Functions such as data deduplication, replication, snapshotting, provisioning, access control, backups and disaster recovery are performed at the SDS level rather than at the device level based on policy-based management directives.

Because many tasks that were previously carried out by IT staff are performed automatically by software, SDS substantially reduces the costs associated with supporting the storage infrastructure.

Standard Interfaces: With SDS all data access and control functions are abstracted from hardware to software. This allows users to interact with storage through a unified “single pane of glass” software interface. Similarly, applications access storage through a standard set of APIs, regardless of the specialized interface protocols and configuration complexities of individual storage units or subsystems.

This simplified data management structure gives SDS a significant ease-of-use advantage over traditional storage.

Virtualized Data Path: Because of its separation of the control and data planes, SDS enables virtual access to storage independent of specific hardware protocols. This allows applications written to block, file, or object interfaces to be natively supported.

Scalability: With traditional SAN arrays, scaling to add capacity can be complex and difficult. SDS, on the other hand, is inherently highly scalable. Software defined storage can quickly, automatically, and transparently scale out to the petabyte or even exabyte range. In the book Building a Modern Data Center, SDS scalability is characterized this way:

“By seamlessly allowing different physical hardware to be added and removed underneath the abstraction layer, changes to the scale of the system can be completed without the workloads even being aware. This gives organizations leveraging SDS a distinct advantage over the prior method of scaling storage…. SDS allows the addition of more storage, or a shift to a new platform to take place, that is totally transparent to the workload.”

Transparency: SNIA defines transparency as “the ability for storage consumers to monitor and manage their own storage consumption against available resources and costs.” A first-rate SDS solution will provide tools that allow visibility into the types, costs, and quality of IT services, both for financial reporting purposes, and to enable adjustments to maximize the efficiency of storage operations.

A Key SDS Feature: Enhanced Flexibility

By their very nature, software defined solutions offer a level of flexibility that’s far beyond what can be attained with traditional storage. Costly proprietary storage appliances and software are no longer required. Instead, SDS software is designed to run on any x86 computeplatform, and can provide sophisticated storage services even with “dumb” commodity hard disks (HDDs) as the storage media. One great advantage of this is that a data center that adopts SDS can often incorporate existing legacy devices and storage subsystems into the SDS solution.

On the other hand, when ease of implementation, operation, and support are the primary considerations, customers can purchase SDS appliances that combine software, server, and storage devices into a single, integrated package. The advantage of this approach is that the SDS vendor can provide comprehensive support services for their product, and the level of engineering expertise required of the customer is minimized.

SDS Is Gaining Momentum

In today’s environment of extremely rapid growth in the demand for storage, many companies have begun to seriously consider how a software defined approach can not only mesh with their present storage infrastructure, but also provide a pathway to the fully software defined data center (SDDC) of the not-so-distant future. As SDS technology matures, and as storage professionals become more familiar with it, the rate of adoption of SDS in both large and small enterprises can only grow.

Want to learn more? Take some time to watch our Talon CloudFAST video.

April 18, 2017

Why SDS Is Becoming More Attractive To CIOs

With the amount of information corporations handle on a daily basis growing at exponential rates, corporate CIOs are having to rethink their approach to data st...
Continue »

April 11, 2017

How SDS Is Disrupting The Traditional Storage Industry

Disrupt: to prevent something, esp. a system, process, or event, from continuing as usual or as expected...
Continue »

March 7, 2017

How to Bridge your Datacenter Needs with Software Defined Storage

Modern business operations are becoming more and more data-dependent every day. The amount of information enterprises must deal with is growing at exponential ...
Continue »