Guest blog article by Eric Barbieri:
In its’ simplest form Software Defined Storage (SDS) can be described as the unbundling of storage software from the hardware necessary to house data. The separation of these two critical pieces of the storage puzzle can have some very profound affects on scalability, flexibility and cost of storing large amounts of data. However, these pro’s much like everything in life aren’t without tradeoffs that should be carefully considered by any enterprise looking to deploy this technology. In this blog, we will explore both the positives and the negatives of this game-changing technology.
First, let’s look at the positives that Software Defined Storage can bring to enterprises of all sizes.
Scalability
Without rehashing the often over used storage growth clichés we all know that we are managing more storage than we even dreamed possible just a decade or two ago. Investing in one of the many enterprise grade storage array platforms often means limitations in the number of disk trays that can be managed by its’ storage processors.
By severing storage software from proprietary hardware-based platforms, Software Defined Storage can often allow storage to be added to pools with sometimes infinite scalability. No more backplane limitations on the number of physical disk drives or throughput limitations on a limited set of physical ports. Software Defined Storage can offer a scale out architecture that increases processing power and overall throughput as nodes are added to overall storage network.
Flexibility
The heart of any Software Defined technology is the ability to install on almost any platform with available resources. While legacy storage platforms have strict requirements on any components that make up a traditional storage array, Software Defined Storage offers the ability to add capacity from many dissimilar sources. Imagine using capacity from an old array, a JBOD of commodity HDDs, Servers with support for 48 NVMe SSD drives all joined together and policed by a software layer completely independent of its underlying technology. The possibilities are endless!
Reduced Costs
While almost obvious at this point, costs can be drastically reduced when deploying large amounts of capacity on commodity based HW. Software Defined Storage can have significant impact on the overall cost of building large scale storage networks.
Now let’s discuss what the potential downsides of deploying Software Defined Storage.
Complexity
Having storage nodes comprised of many different storage types on different hardware platforms may sound like nirvana keeping up with firmware updates and security patches can cause significant headaches as your SDS based network grows. Troubleshooting bottlenecks can also be a challenge when attempting to tune Software Defined Storage based systems for extreme performance.
Standardization
Increased complexity often leads enterprises looking to deploy an SDS based solution to standardize on a standard deployment model. Standardization means limiting choices on HW and often locking in a single or dual vendor’s wares leading to a limit of overall choice and decreasing the option to move to newer technology provided by alternate sources.
Support
Lastly, while nobody enjoys being at the mercy of a single vendor the finger pointing between SDS storage support teams and the underling hardware that makes up the storage network can become troublesome. While traditional storage arrays have their own set of pitfalls its’ often tough to deny the quality of support offered by a large enterprise-focused companies with quality assurance teams that rigorously tested their platforms with the major application platforms. This all can be a major lifeline when the going gets tough.
In closing, SDS offers unique capabilities that many enterprises can utilizes in ways never previously imagined. The maturity of this space has been tremendous and many of the cons of deploying this technology in scale have been ironed out. However, every enterprise must complete their due diligence by evaluating all the potential risks of Software Defined Storage. Lastly, don’t forget to have a properly architected copy data management solution to optimize database cloning, backup and disaster recovery no matter which road you take!